Technical Guide of a Regional Climate Information System

INTERNATIONAL RESEARCH CENTRE ON EL NIÑO - CIIFEN 104

INTERNATIONAL RESEARCH CENTER ON “EL NIÑO”

Escobedo #1204 y 9 de OctubreEdificio Fundación El Universo, 1er piso

Phone: (593 4) 2514770Fax: (593 4) 2514771P.O. Box # 09014237Guayaquil-Ecuador

PROJECT IADB ATN/OC – 10064 – RG

Funded by:Inter-American Development Bank

Regional Public Goods

Executing Agency:

TECHNICAL GUIDEFOR THE IMPLEMENTATIONOF A REGIONAL CLIMATEINFORMATION SYSTEM

APPLIED TO THE AGRICULTURALRISK MANAGEMENT IN

THE ANDEAN COUNTRIES

103 TECHNICAL GUIDE FOR THE IMPLEMENTATION OF A REGIONAL CLIMATE INFORMATION SYSTEM APPLIED TO AGRICULTU-

RAL RISK MANAGEMENT IN THE ANDEAN COUNTRIES

International Research Center on “El Niño” (2009)

It is allowed to reproduce and communicate this guide as long as the source is referenced correctly and it is not used for commercial purposes.

Some copyrightshttp://creativecommons.org/licenses/by-nc/3.0/

Concept, Design & Infographics2009, by Leonardo Briones Rojas

Title pageJosé Benito Valarezo Loor

PhotoAbigail Alvarado, Patricio López, Borja Santos

PrintGráficas Hernández Cía. Ltda.Cuenca, EcuadorDecember, 2009

To refer the whole Technical Guide:Martínez, R., Mascarenhas, A., Alvarado, A., (ed)., 2009. Technical Guide for the Implementation of a Regional Climate Information System Applied to the Agricultural Risk Management in the Andean Countries. International Research Center on “El Niño” –CIIFEN, p 1-160.

To refer one chapter of the Technical Guide: Ycaza P., Manobanda N., 2009. Implementation of Agro-climatic Risk Maps, p 50-62. On the Technical Guide for the Implementation of a Regional Climate Information System Applied to the Agricultural Risk Management in the Andean Countries., Martínez, R., Mascarenhas, A., Alvarado, A., (ed)., 2009. Interna-tional Research Center on “El Niño” –CIIFEN, p 1-160.

ISBN: 978-9978-9934-1-5

This publication was made by the International Research Center for “El Niño” –CIIFEN under the Project ATN/OC 10064-RG “Climatic information Applied to Risk Management in Andean Countries”, funded by the Inter-American Development Bank, IDB, under the initiative of Regional Public Goods (2006).

TECHNICAL GUIDE FORTHE IMPLEMENTATION OF

A REGIONAL CLIMATEINFORMATION SYSTEM

APPLIED TO AGRICULTURALRISK MANAGEMENT IN

THE ANDEAN COUNTRIESPROJECT IADB ATN/OC – 10064 – RG


Executing Agency:

INTERNATIONAL RESEARCH CENTRE ON EL NIÑO

And The National Meteorological and Hydrological Services of

Bolivia, Chile, Colombia, Ecuador, Peru and Venezuela

OCTOBER, 2009

TECHNICAL GUIDE FORTHE IMPLEMENTATION OF

A REGIONAL CLIMATEINFORMATION SYSTEM

APPLIED TO AGRICULTURALRISK MANAGEMENT IN

THE ANDEAN COUNTRIESPROJECT IADB ATN/OC – 10064 – RG


Executing Agency:

INTERNATIONAL RESEARCH CENTRE ON EL NIÑO

And The National Meteorological and Hydrological Services of

Bolivia, Chile, Colombia, Ecuador, Peru and Venezuela

OCTOBER, 2009

Editorial Team:

Rodney Martínez GüinglaAffonso Da Silveira Mascarenhas Jr.Abigail Alvarado Almeida

Project Team:

CIIFEN Staff – Project Counterpart

General CoordinatorRodney Martínez Güingla

International DirectorPatricio López Carmona2006-2007

International DirectorAffonso Da Silveira Mascarenhas Jr.2008-2009

Financial Administration and AcquisitionsRoma Lalama Franco

Geographic Information SystemsPilar Ycaza OlveraMishell Herrera CevallosCarlos Meza BaqueCarlos Zambrano Alcívar

AdministrationMayra Mayorga LópezEvelyn Ortíz SánchezVictor Hugo Larrea Alvarado

Systems EngineeringKatiusca Briones Estebanez

Data and ProductsManagementJuan José Nieto López

Research SystemsAbigail Alvarado AlmeidaAlexandra Rivadeneira Uyaguari

Informatic SupportAlberto Abad Eras

Information AssistantNadia Manobanda Herrera

Risk MapsHarold Troya Pasquel

Regional Agriculturist RiskAngel Llerena Hidalgo

Local Risk ExpertsBolivia Silvia Coca UzunaChile Claudio Fernandez PinoColombia José Boshell VillamarínEcuador Emilio Comte SaltosPerú Oscar Quincho RamosVenezuela Pedro Rodriguez González

Data DigitalizationBolivia José Valeriano Maldonado Luis Bustillos PazChile Viviana Urbina Guerrero Patricia Berrios LeivaColombia Carlos Torres Triana Paola Bulla PortuguezEcuador Carlos Naranjo Silva Ana Zambrano VeraPerú Luis Zevallos Carhuaz Juan Bazo ZambranoVenezuela Vickmary Nuñez Oropeza Gabriel Diaz Loreto

Climatical Data ProcessingPilar Cornejo Rodriguez

Local Experts on Information ManagementBolivia Javier Caba Olguín Chile Miguel EgañaColombia Juan Gómez BlancoEcuador Emilio Comte SaltosPerú Juan Ramos EscateVenezuela Pedro Rodriguez González

Statistical ModelingMarco Paredes Riveros

Numerical ModelingÁngel Muñoz Solórzano

Numerical ModelingRicardo Marcelo Da Silva

Virtual CoreRed de Universidades del Eje Cafetero Alma Mater

Data BaseCentro de Tecnologías de la Información - ESPOL

NATIONAL METEOROLOGICAL SERVICES

BOLIVIA

DirectorCarlos Díaz Escobar

Statistical ModelingGualberto Carrasco Yaruska Castellón Nidia ZambranoVirginia Rocha

Dynamic ModelingGualberto Carrasco Erick Pereyra Ramiro Solíz

Agro-Climatic Risk MapsYaruska CastellónOscar Puita

Focal Point for the ProjectPablo Elmer

METEOROLOGICAL AND HYDROLOGICAL NATIONAL SERVICE - SENAMHI www.senamhi.gov.bo

CHILE

DirectorMyrna Araneda Fuentes

Statistical ModelingJuan Quintana

Dynamic ModelingClaudia Villarroel Roberto Hernández

Information SystemsMiguel Egaña

Agro-Climatic Risk MapsPatricio LucabecheJosé Curihuinca

Focal Point for the ProjectGualterio Hugo Ogaz

METEOROLOGICAL DIRECTION OF CHILE - DMC www.meteochile.cl

COLOMBIA

DirectorCarlos CostaRicardo Lozano

ModelingGloria León Aristizábal

Quality AnalysisRuth Correa Amaya

Agro-Meteorological AnalysisGonzalo Hurtado MorenoRuth Mayorga Márquez

Focal Point for the ProjectErnesto Rangel MantillaChristian Euscátegui

INSTITUTE FOR HYDROLOGICAL, METEOROLOGICALAND ENVIRONMENTAL STUDIES - IDEAM

www.ideam.gov.co

ECUADOR

DirectorCarlos Lugo

Statistical ModelingCristina Recalde

Dynamic ModelingJaime Cadena

Agro-Climatic Risk MapsFanny Friend

Focal Point for the ProjectRaúl MejíaFlavio Ramos

NATIONAL INSTITUTE OF METEOROLOGY AND HYDROLOGY - INAMHI www.inamhi.gov.ec

PERÚ

DirectorWilar Gamarra Molina

Statistical & DynamicModelingCarmen Reyes Bravo Juan Bazo Zambrano

Information SystemsLuis Zevallos Carhuaz

Agro-Climatic Risk MapsDarío Fierro ZapataKevin Sánchez ZavaletaNelly Perez Díaz

Focal Point for the ProjectDarío FierroConstantino Alarcón

NATIONAL SERVICE OF METEOROLOGY AND HYDROLOGY - SENAMHI www.senamhi.gob.pe

VENEZUELA

DirectorRamónVelásquez Araguayan

Statistical & DynamicModelingLuis MonterreyAlexandra MataElddy Anselmi

Data Systemsand DigitalizingRichard NúñezJenny CastilloManuel González

Agro-Climatic Risk MapsCarlos OjedaLuis MonterreyCésar Yauca

Focal Point for the ProjectAlexander Quintero

METEOROLOGICAL SERVICE OFTHE BOLIVARIAN NATIONAL AVIATION

www.meteorologia.mil

NATIONAL METEOROLOGICAL SERVICES

Since it was established in January 2003, one of the most important mandates of the International Research Center on El Niño (CIIFEN) has been to build the

necessary bridges between the climate informa-tion suppliers and the users from other sectors of the society.

The ultimate goal is to use the benefits provided by earth observation, science and predictions in order to allow our society to live better. When

we talk about risk management, it results in reduced loss of life and goods for development support.

To review all the climate information to make it a tool for human welfare is not easy, it requires a holistic vision, inter-and trans-disciplinary dialogue and, above all, it requires breaking several paradigms.

In 2003 the World Meteorological Organization, through its Division of Services and Climate Applica-tions, organized, in alliance with CIIFEN, a regional workshop to identify the needs of climate informa-tion for the agricultural sector. This meeting provided us with essential information to generate (after several years) a regional proposal that addresses the needs of this important sector.

In 2006, the Inter-American Development Bank (IADB), under the modality of Regional Public Goods, approved the project entitled “Climate Information Applied to Agricultural Risk Management in the Andean countries” to be carried out by CIIFEN and the National Meteorological Services of Boli-via, Chile, Colombia, Ecuador, Peru and Venezuela.

After three years of efforts, regional cooperation and with the trust and support of the IADB, we can bear witness to this important initiative through this Technical Guide, which describes step by step how we implemented the system in each one of its components, including the learned lessons, sustainability strategies and future challenges.

With deep gratitude to the Inter-American Development Bank (IADB)The World Meteorological Orga-nization (WMO), The National Meteorological and Hydrological Services (NMHSs) and the Internatio-nal Research Centre on El Niño (CIIFEN), we present the “Technical Guide for the Implementation of a Regional Climate Information System Applied to Agriculture Risk Management in the Andean Countries”. We hope that it can be replicated in other places around the world for the benefit of our society.

Dr. Affonso MascarenhasInternational Director

CIIFEN

Oc. Rodney Martinez GüinglaProject CoordinatorATN/OC 10064-RG

INTRODUCTION


Introduction

Chapter I: Implementation of the Virtual Core for Climate Applications (VCCA)

1.1 Conceptual model

1.2 Technologic Platform

1.2.1 VCCA Architecture 1.2.2 Physical Infrastructure 1.2.3 Logical Infrastructure

1.3 Applications That make up the VCCA

1.3.1 Regional Climate Data Base 1.3.2 Map Server 1.3.3 Display of Climate Model Products 1.3.4 Virtual Library

1.4 Implementation Process of Regional Climate Data Base

Chapter II: Implementation of statistical modeling for climate prediction

2.1 Conceptual and methodological elements 2.2 Management to update the information of the Predicting Variables

2.2.1 How to perform the Alternative Method for Forecast Updates.

2.2.1.1. Procedure for obtaining the sea surface temperature (SST) variable. 2.2.1.2 Procedure for obtaining the wind variable at altitude, geo-potential, temperature at mandatory levels.

2.3 Operating simultaneous predictors with CPT.

2.3.1 Climate Forecasting with Simultaneous Predictors

2.4 Decision criteria for managing the CPT results. 2.5 Considerations for the interpretation of terciles 2.6 FAQs related to CPT operation.

Chapter III: Implementation of numerical models for climate prediction

3.1 Step by step installation and implementation procedures for MM5 and WRF models in Climate Mode.

3.1.1 Operating System 3.1.2 Atmospheric Models

3.1.2.1 MM5 3.1.2.2 CMM5 3.1.2.3 WRF 3.1.2.4 CWRF

107

111

112

112

112112112

113

113114115115116

117

118

119

119

119

121

122

122

123

127

127

131

132

132132

132137137139

INDEX



3.1.3 Oceanographic Models

3.1.3.1 ROMS

3.1.4 Displayers

3.1.4.1 GRADS 3.1.4.2 Vis5D

3.2 Implementation of Numerical Modeling for Climate ForecastingThe Regional Group of Numerical Modeling

Chapter IV: Implementation of Agro-Climatic Risk Maps

4.1 Definition of Risk

4.2 Conceptual Mathematical Model of Agro-Climatic risk

4.3 Components and Variables of Agro-Climatic Risk

4.3.1 Thread 4.3.2 Vulnerability

4.4 Project Application Areas

4.5 Information Requirements

4.5.1 Agro-ecological 4.5.2 Base mapping 4.5.3 Thematic mapping 4.5.4 Treatment of information 4.5.5 Soil and climate characteristics in pilot areas

4.6. Agro-Climatic Risk Calculation

4.7. Agro-Cimatic risk in the Andean Countries

Chapter V: Implementation of local systems of climate information

5.1 Conceptual and methodological elements.

5.2 Identification and mapping of key components.

5.3 Strategic Alliances

5.4 Strategic alliances with local authorities.

5.5 Strategic alliances with the private sector. 5.5.1 Journals Specialized in Agriculture 5.5.2 Cell phone Company

5.6 Strategic alliances with the media.

5.7 Training Strategies

140

140

141

141141

143

145

146

146

146

147147

148

150

150150150150150

152

154

161

162

163

164

165

165

166167

168

169


Chapter VI: Capacity building in Western South America

6.1 Regional Training Workshop on Climate Modeling statistics

6.2 Regional Training Workshop on Numerical Modeling for Climate predictors

6.3 Regional Training Workshop for Agro-Climatic Risk Mapping

6.4 Regional Workshop on Numerical Modeling of Weather and Climate II

6.5 International Training Workshop on Climate Data Processing

Chapter VII: Performance Indicators.

Chapter VIII: Learned Lessons

Chapter IX: Future Actions

Chapter X: Elements of Sustainability

Bibliographic References

173

174

174

174

175

175

177

181

185

187

189



CHAPTER IImplementation of the Virtual Core for Climate Applications

(VCCA)

CHAPTER I

1.1 CONCEPTUAL MODEL

The Virtual Core for Climate Applications (VCCA) consists of a basic computing infrastructure to run climate applica-tions designed to provide climate information through an easy to use and easy access format through the Internet.

Under this philosophy, the VCCA centralizes all the neces-sary functionalities for different types of advanced WEB ap-plications, such as: products presentation, users’ control, management of geographical information system, biblio-graphic information and search for information.

1.2 TECHNOLOGIC PLATFORM

1.2.1 VCCA Architecture

The main purpose of the applications running on the VCCA is to provide information to end users, without requiring the installation of any special software. This established the client-server architecture, in which CIIFEN would be in charge of the central server, and the end users would have access through web interface using the Internet. This allows simultaneous user connectivity and protection of published information.

Figure 1 shows VCCA architecture graphically, in which the physical infrastructure (servers), the logic infrastructure (software) and the end users intervene.

1.2.2 Physical infrastructure

The VCCA was implemented with two servers, for the ad-ministration of CIIFEN’s internal network and VCCA instal-lation. The servers used are: Dell PowerEdge 2950 with a Xeon Dual Core 2.66GHz processor with 4GB of memory and a disk capacity of 600GB (primary server), 300GB (sec-ondary server) and RAC type servers.

The technical details of the software used in the VCCA are described as follows:

Operating SystemsManagement and applications servers, run on Linux SUSE Operating System V.10, which has been demonstrated to be sufficiently stable, ensuring the availability of permanent climate applications over time.

Database Management System (DMS)Database Management Systems DMS are intended to sup-port the tasks of defining, creating and manipulating re-lational databases; for which that allows for features such as: concurrency control, data backup methods and access control applying user profiles.

The virtual core operates with the DMS, called PostgreSQL 8.3; this is a Relational Object type system, and is used extensively due to the characteristics of the standards ap-plied, the securities and the capability of communicating with various types of applications, among which is the abil-ity to store spatial data, which is needed for applications of the geographical information systems.

Spatial Information Management System (SIMS)The web visualization of cartographic information, agricul-

Figure 1. Virtual Core for Climate Applications Architec-ture (VCCA)

Figure 2. Software Infrastructure of Virtual Core for Clima-te Applications (VCCA)

1.2.3 Logical Infrastructure

The client-server architecture requires a server capable of conducting central processing of the applications running on it, while clients “ask for” information without having to process it internally. On this logic infrastructure scheme, a centralized databases is created according to the applica-tion, so the information is kept in one place with the partic-ularity of being accessible for viewing and/or maintenance, depending on the type of user (end user or administrator).

The developed applications communicate with the cor-responding database independently, placing the visual interface over the one that is displaying the information re-quested by the user (Fig. 2).

Katiusca [email protected]



C H A P T E R I

ture climate risk and geographic information system re-quires the use of several special-purpose tools that togeth-er allows the full functionality of the map display system. The tools used are:

PostGIS: Modules under GNU license provide the Post-greSQL Database management system with the ability to manage spatial information.

MapServer 5: CGI Aplication (Common Gateway Interface) which is a standard for communicating between a Web server and a program, so that the user can interact through the Internet (in the case of dynamic maps).

Grass: Geographical Information System, that handles the information on the web.

Web Server Apache 2.0: HTTP SERVER (the protocol that defines the semantics used for clients and servers to communicate with each other); it is multi-platform open code. Its architecture allows the addition of modules to provide several functions, such as dynamic webpage support and message encryp-tion.

Application SupportJava Application Platform (SDK): Platform on which certain components (climate database) of the VCCA are run.

Perl: Program to run certain components of the applica-tions (display of numerical modeling products). End UserOne of the goals outlined in the development of VCCA, was to eliminate the need for the user to install any spe-cial software. To access any of VCCA applications, the user needs only to have an Internet connection, run his preferred browser program and access the appropriate link.

1.3 APPLICATIONS THAT MAKE UP THE VCCA

The project developed the applications in the VCCA:• Regional Climate Data Base: http://vac.ciifen-int.org• Map Server: http://ac.ciifen-int.org/sig-agroclimatico • Display of Climate Modeling Products:http://ac.ciifen-int.org/modelos • Virtual Library: http://ac.ciifen-int.org/biblioteca/

1.3.1 Regional Climate Data Base

The Regional Climate Data Base corresponds to an appli-cation for displaying climate data of temperature and pre-cipitation in the Andean countries (Bolivia, Chile, Colom-bia, Ecuador, Peru and Venezuela).

The Regional Climate Data Base for Western South Ameri-ca, is an cooperative effort without precedents among the National Meteorological Services of the region and is a gi-ant step towards the integration of climate data to be used use in regional forecasting and also to contribute to atmo-spheric sciences research.

This information resource is made possible thanks to the unflinching support and hard work of the NMHSs of Bolivia, Chile, Colombia, Ecuador, Peru and Venezuela. The data-base is one of the columns of the climate information sys-

tem applied to agriculture climate risk management in the Andean countries as a Regional Public Good that contrib-utes to the understanding of past climate and its important role of evolution over time.

Figure 3 shows the Database which is available at http://vac.ciifen-int.org and contains records from 170 meteorological stations from 1952 until 2008 and is the start of a data ex-change system without precedent, which will also improve climate forecasting services in the region. It contains daily records of Precipitation, Maximum Temperature, Mini-mum Temperature, Basic Data Stations and also displays climate products as time series or spatial graphics.

Figure 3. Startup Screen of the Regional Climate Database

The application allows to view different types of graphs (time series, contour plots, histograms) and to check me-teorological stations (location, general information). For the creation of the Databse and its updating, a Protocol between National Meteorological Services and CIIFEN was signed (Annex I).

Available chart types:The application provides three sets of information:

Data Search: allows us to select the graphic display of time series and histograms, and also to download the data in text format of the CPT model1 in maximum / minimum / accumulated monthly, bimonthly, quarterly or yearly. (Fig. 4) (Fig. 5)

1. Climate Prediction Tool, http://portal.iri.columbia.edu

Figure 4. Screen of Data Search by Regional Climate Database Stations


C H A P T E R I

Figure 6. Detail display screen of the Regional Climate Database Stations

Figure 7. Selection screen of Climatic Products of the Regional Climate Database

Figure 8. Graphics screen of Climate Products of Regional Climatic Database by country

Figure 5. Display Screen of Time Series of the Regional Climate Database

Stations: Displays a list of all the weather stations involved in the ATN / OC 10064-RG project, identifying details of its basic information, location and additional information for each of these. (Figure 6)

Climate Products: Displays spatial graphics using a for-mat of isolines for precipitation and temperature, in which is possible to select an area of a country or of the South American Continent (Figure 7) (Figure 8).

Data UpdateThe Climate Database application is fully upgradeable; to allow this it has an administrator interface, in which each country can connect through the same interface and up-load the data files.

1.3.2 Map Server

The Map Server Application aims to provide the user with the ability to visually manipulate different levels of GIS in-formation through a friendly web interface, without running any specialized software in the computer.

This Web-based interface provides the ability to view any fi-nal product of a GIS, as is the case of the agriculture climate Risk GIS, the initial product placed on the display. We need to indicate that a pre-processing of the levels is necessary in order to publish them from shape to xml format.

Graphic InterfaceThe graphical interface of the Map Server allows end user to select the different Andean countries involved in the project. For each one of them, available information is dis-played. (Figure 9).

Figure 9. Map Server Startup Screen

Each link within countries displays a list of layers and topics developed in the project. The selected layers and selected topics are displayed in a GIS management interface, in which the user can hide/display layers, zoom in/out, display information from the components of each layer, select com-ponents, measuring tools, insertion of points of interest,



C H A P T E R I

Figure 10. Map download tool in GeoTIFF format from the geographic information system based on the Web.

1.3.3 Display of Climate Model Products

The goal of the Climate Modeling Products display is to create an application on which the end user can display products from different numerical climate models (Fig 11).

Figure 11. Home Screen Numerical Modeling Products Display

The developed Web interface allows the user to choose the climate model he wants to view and select the dates on which forecasts have been made. Once the user has chosen the date, he can select the domain and the climate variable, and then, the corresponding product will be dis-played. (Figure 12)

The interface on which numerical modeling products are published is Google Earth satellite images interface, which makes this application a topographic information tool that is allows to visualize areas of different altitudes when ana-lyzing climate forecasts. (Figure 13)

1.3.4 Virtual Library

The purpose of the Virtual Library is the systematization of the vast amount of information that CIIFEN has collected

Figure 12. Display screen of climate model products, Accumulated Precipitation variable

Figure 13. Display Screen Climate Model Products, Air Temperature variable

and download images in GeoTIFF format (geo-referenced image). (Figure 10)

The application has been developed for the user to load new layers of information, for which it is necessary to trans-form each layer from shape format to xml format.

Figure 14. Startup screen Digital Library

since its creation, the goal of the application is to publish books, magazines, reports, presentations, CDs, and more free access information sources and to disseminate it to the general public. (Figure 14)

The virtual library is published at: http://ac.ciifen-int.org/biblioteca. There are two search options: by Books and Digital Archives:

Books Section• Contains information of books, magazines, journals, at-lases, and other paper publications.


Figure 15. Display screen Digital Library publications

Administrative interfaceThe application has an administrative interface through the Control Panel option, in which the library administrator has several management options, such as adding categories, add/edit/delete/reserve publications and add/delete us-ers.

1.4 IMPLEMENTATION PROCESS OF A RE-GIONAL CLIMATE DATA BASE

Integrating climate information from the Andean countries was a fully coordinated joint effort in which the NMSs pro-vided maximum collaboration to the compilation of nation-al databases. This process was executed in five stages:

• Collection of information: In order to determine the availability of information in different existing formats with-in each National Weather Service, they proceeded to per-form a survey on the staff, which determined the amount of digital data and hard copy information

• Acquisition of computer equipment: The digitization of information required the acquisition of computer equip-ment; two computers for each NMHS were designated for this purpose.

• Hiring of operators: The amount of information to be loaded was based on surveys, and it was coordinated with each NMS to hire two operators for each institution. They processed the information in the appropriate formats.

• Compilation of information: The digital information was added to the databases of each NMS, increasing the den-sity of climate data in each institution.

• WEB Application Development: Based on information gathered by each NMS, the Web application was devel-oped with the data of precipitation, maximum and mini-mum temperature. In the application, a database mainte-nance module was developed in which a representative of

each country can access using the appropriate username and password to manage their data and also to add new information.

• The Regional Climate Database contains 3,876,035 cli-mate records and will maintain up to date according the established protocol.

C H A P T E R I

Digital Archives Section• Contains information of presentations, CDs and DVDs of applications, reports, data and projects which CIIFEN has compiled from the various events in which it has participat-ed. These archives are available for free.

The application has a search interface in which the user can enter keywords and select the search type. As a result, it will show all matches found in the library, identifying the information of each publication. (Figure 15)



CHAPTER IIImplementation of statistical

modeling for climate prediction

2.1 CONCEPTUAL AND METHODOLOGICAL ELEMENTS

The tool used for the implementation of statistical mod-eling in every country was the Climate Predictability Tool (CPT), developed by the IRI. The flow chart of actions de-signed for the region is explained in the Figure16.

Figure 16. Process for the realization of the seasonal forecast, using the CPT Figure 17: Pressure Monitoring conducted at sea level by

NOAA / NCEP / NCARFor the purpose of using the CPT, the information from the National Meteorological Services is collected on the 30th of each month (28th in February), before the update done by the Global Forecast Center. This will be basic information that serves as a predictor, under the following assumptions:

It shows the average to date. In the right corner it shows the anomalies that occurred during that period and at the bottom the climate for the same period, to compare them.Likewise, we should keep in mind the monitoring of SST

The Sea Surface Temperature (SST), due to variable inertia, experiences slowly changes in its physical patterns. Under this premise, any change in the next four or five days will not be significant on the monthly average, which is carried out by averaging the first 27 days of the current month and is attached to the time series of sea surface temperature that can be obtained from NOAA/NCD/ERSST. At the end a complete and updated historical series is obtained, which serves as the final predictor.

Atmospheric variables, such as zonal wind, southern wind, temperature at high levels, specific humidity, among oth-ers, we must take with extreme caution the changes in these last five days of the month. They can be significant and may modify the average. It is therefore advisable to monitor climatic conditions globally and in particular South America or the region of interest.

The analysis of the monitoring of various oceanic and atmo-spheric variables should be conducted every two weeks. If possible, it is recommended to do it on a weekly basis, as shown in Figure 17:

Marco [email protected]

CHAPTER II

Development ofCPT Data

Regionalizationof variables

Climatemonitoring

Updating ofpredictors

Regionalizationof predictor

Runof CPT

Analysisof results

Assemblies

Spatial Graphics

Dissemination



and its anomalies and their influence on changes in climate patterns.

2.2 MANAGEMENT INFORMATION UP TO DATE FOR VARIABLE PREDICTORS

2.2.1 How to perform the Alternative Method for Forecast Updates

The monthly predictive indices of global forecasting cen-ters provide information that is available in the IRI’s Data Library. They are updated on the 10th of each month with information from the previous month. As a result, the statis-tical model run are delayed until a date after the 10th for SST and the 15th for other predictors.

To avoid this we have chosen an action that allows an up-date a few days in advance of some of the necessary pre-dictors, especially sea surface temperature (SST) under the following circumstances:

• It is considered that the predictor to be analyzed will not undergo significant changes when it is completed with the missing data until the end of the month.• 75% of the days of the month are averaged so as to be considered representative. This means that at least 21 days of the month must have already passed.• The changes in the SST’s values have no abrupt behavior because of the ocean’s inertia (specific heat, which allows a delay in heat loss by up to 5 times longer than on land).

2.2.1.1 Procedure for obtaining the sea sur-face temperature (SST) variable

In this case, there must be prior historical information from the SST variable obtained from NOAA / NCDC / ERSST in order to predict the “Y” variable.

For weekly SST data, the corresponding search in the IRI data library is performed in the Air-Sea interface category. (Figure 18).

The data belongs to the NOAA / NCEP / EMC global CMB, Reynolds Smith research center. It should be looked for weekly data by entering to version 2 of the Reynolds data (Reyn Smith IOv2). (Figure 19)

Press the weekly data (weekly): (Figure 20).

Figure 18. Air-Sea interface Data in the IRI Data Library.

Figure 19

Choose to download the data of the weeks of interest (Fig-ure 22).

The process is the same that was used to download infor-mation on any other variable. The only difference occurs in the window Time (symbolized by the letter T). It places the weeks of the month, considering that it begins Sunday and ends on a Saturday, for example, for the weeks of February are: (Figure 23)

Choose the option Sea Surface Temperature (SST) (Figure 21).

Figure 22

Figure 21

Figure 20

Figure 23

C H A P T E R II


Download the monthly average data.

Note: Keep the following in mind before the process: the resolution of the Reynolds data is 1° x 1° and is not com-patible with the data of CPT when it runs (Source: NOAA / NDCD / ERSST whose resolution is 2° x 2°). To solve this incompatibility a spreadsheet has been compiled (called transformation) which converts the Reynolds data to ERSST data.

The chart below shows the format obtained through the process described above, where the first line and first col-umn indicate resolutions in longitude and latitude respec-tively (1° x 1°). (Figure 25)

Copy from the second line all the obtained information in the file and take it to the spreadsheet 1 of the file TRANS-FORMATION, on the yellow background area (copy and paste), leaving the first row empty. (Figure 26).

You get 03 sets of data (03 weeks); therefore you should perform filtering of information with an average time (called T in CPT), which shows the average of 03 weeks elapsed, corresponding in this example to February. From this part, there are two ways to standardize the resolution between the two data sets from different research centers, which are detailed below:

First way: (Figure 24)

Figure 24

Figure 26

Figure 25

Then, enter on page 2 of the same file and save as text file (which is already transformed). (Figure 27)

The format obtained is the following: (Figure 28)

Save it to add it to the February predictor history, with a single copy. (Figure 29)

Figure 27

Figure 28

Figure 29Then record it.

C H A P T E R II

PASTED FILE(Weekly averages in the interest month)



Note that will have information, it starts in one year 1960 and ends with the year 2008, which is ready for inclusion on the CPT. (Figure 30)

Second way:

Login to expert mode after obtaining the average week (found with a resolution of 1° x 1°), perform the following commands.

X 0 2 358 GRID

Y -88 2 88 GRID

Then download it, save it and paste it on the base obtained from the initial historical series (procedure described in item 2.3.1.1), which is ready for use as a predictor.

2.2.1.2 Procedure for obtaining the wind vari-able at altitude, geo-potential and tempera-ture at mandatory levels

If is the case to work with an atmospheric variable at alti-tude, there is a practical procedure to work with averages of days elapsed. An option is to use information NOAA NCEP-NCAR CDAS-1 which lies within the model simula-tions (HISTORICAL MODEL SIMULATIONS). (Figure 31 and 32).

Figure 30

It is necessary to previously have monthly historical data of the variable of interest from the same research center (NOAA NCEP-NCAR CDAS-1) so they have the same reso-lution in order to fit together more easily. (Figure 33)

Choose the daily data (DAILY) and subsequently the IN-TRINSIC mode. (Figure 34)

Figure 32

Figure 33

Figure 31Figure 34

When information regarding altitude level is required, choose the Pressure Level option, which lets you choose the level of interest.

Variables that can be provided are multiple; however, the most common are: geo-potential height, zonal wind, south wind and temperature. They can be seen in the following screen (Figure 35)

C H A P T E R II

Delete date dashes by blank spaces

Look for resolution changes in first row and column: 2º x 2º.Delete date dashes by blank spaces


Figure 37

Figure 35

Figure 36

The available levels in the library are 1000, 925, 850, 700, 600, 500, 400, 300, 250, 200, 150, 100, 70, 50, 30, 20 and 10 mb. (Figure 36)

Perform the filtering with T (average time due to the avail-ability of 22 series, one for each day) and then proceed to download data.

Note: for these variables it is not necessary to change the resolution, or use the TRANSFORMATION spreadsheet. (Figure 37)

2.3 OPERATING SIMULTANEOUS PREDIC-TORS WITH CPT

2.3.1 Climate Forecasting with Simultaneous Predictors

The run with each predictor must be done individually to obtain the weights that influence the variable to be pre-dicted with some additional comments.

Place the maximum number of modes for the variable X, which corresponds to fewest number obtained between the number of years in historical series and the number of grid points or seasons. The maximum number of variable X is required, so you take 43 (according to the example), although in reality the maximum number for X is 44 (n final - initial n + 1), only that the ordered common pair 1965-2007 is considered so there is an additional process for calculat-ing the weight of 2008, which is described below (Figure 38).

Figure 38

You should obtain from the menu FILE/OPEN FORECAST, and then place the start year and number of end years (in-cluding 2008). (Figures 39 - 40)

Figure 39

Figure 40

C H A P T E R II



Display the forecasts through the menu: FILE/FORECAST/SERIES (Figures 41-42)

Continue with the same procedure for second or more vari-ables (or the second area as the case may be) and then group into a single file, which will act as a predictor for the variable under discussion (Figure 44).

And you obtain the file under the following format: (Figure 43)

Figure 44

Figure 43

Figure 42

Figure 41

In the CPT, the forecasting process for the year of interest is the same as is performed with individual predictors, by placing the X EOF option in MATRIZ VARIANZA–COVARI-ANZA (MATRIX VARIANCE – COVARIANCE) to preserve the relative importance of the EOF (Figures 45-46-47).

Figure 45

Figure 46

Figure 47

2.4 DECISION CRITERIA FOR MANAGING THE CPT RESULTS

1. One of the first indicators to be displayed is the GOOD-NESS INDEX which is the result of the first interaction be-tween the predictors and predicting variables; this is the first condition to follow. If we obtain a negative value, it in-dicates no correlation or linearity between the information from both variables; for this we must find a better area. Pref-erably, this value should be positive and higher (tendency to have a value of 1). (Figure 48)

C H A P T E R II


Figure 48. Display in GOODNESS INDEX

2. One of the most important criteria to be considered in the forecasts is the definition of the work period to be used, which is defined by two things:

• The “LENGTH OF TRAINING” PERIOD and,• The “FIRST YEAR OF X TRAINING” PERIOD.

3. In the definition of the climatological period to work on, usually the program defines by default the start and end years of the historical series (in many cases exceeds 30 years). When considering different periods, there will be different results.

The climatologic reference period considered was 1971-2000; many researchers considered the normal since the start of the historical series until the year preceding the forecast.

The change can be accomplished through the following steps: Enter the CUSTOMIZE menu (configuration), then “Climatological Period”. (Figure 49)

And then modify the years. (Figure 50)

4. Check the canonical correlation coefficient, which is the degree of relationship between the predictor and predic-tant variables (jointly). (Figure 51)

First run of the CPT program, where the index is located.

5. Only if the previous step is satisfactory, proceed to evalu-ate the statistical indicators through individual assessment by station, considering the following route:

TOOL/VALIDATION/CROSS VALIDATED/PERFORMANCE MEASURES /

The analysis is done station by station; at this stage you can not see the stations that exceed the permissible limit of missing data (% MISSING VALUES) (Figure 52)

First, display the chart and compare the red line (observed values) and the green lines (forecast values); highlight if the curves follow the same characteristic pattern, i.e. if a curve rises, the other has to rise and vice versa.

The second display is done in the ROC graph (Relative Op-erating Characteristic) where you can see the curves that are found above the diagonal. If the curve is red, it refers to the predictions made by the model to the “below normal”

Figure 50

Figure 51

C H A P T E R II

Figure 49



category; if it is blue, it refers to the predictions made by the model in the “above normal” category. It is appropriate that the two curves be above the diagonal and approach-ing the upper left corner.

6. Second step, although the statistical indicators are a technical reference, you should fully understand their meanings. The first coefficient of Pearson’s1 and the one of Spearman2 indicate the degree of association that the observed values have with forecast values, and should be approximately 1; the higher these values are, the more fa-vorable the results will be (not good to get values close to -1). (Figure 53)

Figure 52

Figure 53

The mean squared error and root mean squared error have the same meaning: they represent the sum of deviations between observed and forecast values, i.e. the error that exists so that predicted values to try to reach the observed value. In a practical way, if the observed and predicted val-ues are similar or nearly identical, means that the error will be zero or nearly zero, hence also its square root. It should be considered that this indicator is very relative: it is not the same to find a difference between both values (observed and predicted) in a rainy area than in a dry area, for example:

1. Randall E et al. A beginner’s guide to structural equation mode-ling pg. 38.2. William H. Press. Numerical recipes: the art of scientific compu-ting pg. 349.

7. The other values refer to categorical measurements, i.e. the level of accuracy of the model with historical data.

Hit Score: The percentage of hits of the model relative to total predictions made from historical series.

The optimum is to have a value close to 100% which would indicate a perfect model.

Hit Skill Score: Is the indicator for evaluating the skill of the model, the percentage of times the result corresponds to a coincidence. The optimum is to have a value close to ±100% which would indicate a perfect model.

LEPS Score (Linear Error in Probability Space), That cal-culates a defined result using a table that shows different results of accuracy, depending on the category observed and the previous probabilities of the categories. The prob-ability distribution is transformed to a cumulative probabil-ity function. (Figure 54)

Gerrity score: Calculates a definite result by using a result table alternative to that used for LEPS results. (Figure 55)

Figure 55

Figure 54

First Step

Good forecast

Bad area

C H A P T E R II

Forecast

Precipitation

Observed

Precipitation

Error Observations

430 mm / month 380 mm / month 50 mm Wet zone 10 mm / month 0.0 mm / month 10 mm Dry Zone

Forecast

Precipitation

Observed

Precipitation

Error Observations

430 mm / month 380 mm / month 50 mm Wet zone 10 mm / month 0.0 mm / month 10 mm Dry Zone


Scheme N°01.- Previous processes for the run of the Climate Predictability Tool (CPT)

Scheme N° 02. - Process Assessment and De-cision Making of the results obtained by the CPT

Nota: The symbol means tendency or approach

ROC area (below-normal): Represents the value of the area below the red curve. Defines the area below the ROC curve for forecasts of the below normal category; shows the proportion of times that below normal conditions can be successfully distinguished over other categories. A maxi-mum value and optimal in the model should be 1 (meaning 100%).

ROC area (above-normal): Represents the value of the area under the blue curve. Defines the area below the ROC curve for forecasts of the above normal category and shows the proportion of times above normal conditions can be successfully distinguished over other categories. A maxi-mum value and optimal in the model should be 1 (meaning 100%).

Run of CPT

C H A P T E R II

Entry formats of the CPT: Quarterly, Bi-monthly, Monthly

Determination of new variable and/or

predictor area

Initiate alternative process of variable

update

Start year, missing data.

If the variable = PP-Y bound=0

Number of years = total of the serie

Is the predictor updated?

Ready data

Configuration of the CTP

Number of modes:X=10Y=10

CCA=10

NO

YES

Goodness index -1

no

yes

yes

yes

yes

yes

no

no

no

no

no

Joint Evaluation

CCA-1/high?

Evaluation of models for

each station

Graphic evaluation. Observations vs

Similar forecasts?

Graphic evaluation. ROC: above the

diagonal?

Evaluation of Good Categorical

measurements

Use for individual forecast



If the requirements of Schedules 1 and 2 are met, we are able to use the model to forecast the year preceding each station (individually) that met all these requirements. For this, we carry it out through the menu: (Figure 56)

Place the year to be forecast: 2008 (first year of data in file). (Figure 57)

In the menu: TOOL/FORECAST/MAPS/PROBABILITIES.

In the probabilistic outcomes only the seasons that met all as indicated in Figures 1 and 2 should be considered. The rest of the values will not be considered for making the forecast table and will be determined with other indexes. (Figure 59)

2.5 CONSIDERATIONS FOR THE INTERPRE-TATION OF TERCILES

The CPT considers among its results by categories values above 50% as extreme (superior and inferior). The value of normal condition is the same as saying the likelihood of the climatology. For example, the following graph shows

Figure 56

Figure 59

Figure 60

Figure 57

Figura 58

probabilistic values of an above normal condition (supe-rior). (Figure 60)

If you have values below 50% in category B (below normal) and A (above normal), they are considered normal, as for example, a probability of 25 - 30 - 45, for the CPT to be con-sidered very close to the upper limit but within the “Nor-mal” category. It is noteworthy that many researchers find no significant differences between the values of 25-30-45, considered as either 03 possible cases.

2.6 FREQUENTLY ASKED QUESTIONS RE-LATED TO CPT HANDLING

1. What do I do if one of the requirements of Schedule 1 and 2 is not met?

In that case you should discard the values of the station; therefore, it is not considered in the final results.

2. How do I consider in the event that the CCC is favor-able and in the individual analysis by stations only a few are favorable?

In that case, only those that are both favorable to the ca-nonical correlation coefficient (CCC) and individual station statistical indicators will be considered in the eventual out-come.

C H A P T E R II

50% Limit

ABOVE

BELOW


5. Does the CPT provide deterministic values in its fore-casts?

The CPT has the advantage of performing multiple opera-tions; therefore, it provides multiple results: one is estimat-ing the values of quantitative forecasts under a given confi-dence level (by default the program calculates with a 68.3% confidence level).

This can be displayed after enabling the forecasting by series, following the completion of the run: TOOL/FORE-CAST/SERIES / (Figure 62)

6. What is meant by confidence intervals?

Interpretation 1A confidence interval is a range of values that has a given probability of containing the parameter being estimated. The 95% and 99% confidence intervals, which have a 0.95 and 0.99 probability of containing the parameter, respec-tively, are the most used.

If the parameter being estimated were m, The confidence interval of 95% will be:

12.5 ≤ m ≤ 30.2

Figure 61

Figure 62

This means that the interval between 12.5 and 30.2 has a 0.95 probability of containing m. We can also say that if the procedure for calculating the confidence interval of 95% is used many times, 95% of the time the interval will contain the parameter.

Interpretation 2It is called confidence interval in Statistics to an interval of values around a sample parameter where, with a deter-mined probability or confidence level, the population pa-rameter to be estimated will be situated. If α is the random error that you want to assign, the probability will be 1 - α. At a lower level of confidence, the interval will be more precise, but it will commit a greater error.

To understand the following formulas, it is necessary to un-derstand the concepts of parameter variability, error, confi-dence level, critical value and α value.

A confidence interval is thus an expression of the type [θ1, θ2] or θ

1 ≤ θ ≤ θ2, where θ is the parameter to be estimated. This interval contains the estimated parameter with a given certainty or confidence level 1-α.

Upon providing a confidence interval, it is assumed that population data are distributed in a certain way. It is cus-tomary to do so by normal distribution. The construction of confidence intervals is performed using the Chebyshev inequality. (Figure 63)

3. How do I consider if the Pearson and Spearman coef-ficients are high but negative?

They are not considered in the analysis. The values by sta-tion are discarded and not considered in the final grouping of forecasts.

4. How do I obtain the limits of the climatology values?

There are two ways to get the climatology:

The first comes from the same original data (data format of the CPT entry corresponding to the variable you want to predict= Y); you should add to each column the percentile values 33 and 66, which corresponds to the limits of the ter-ciles. This value is variable depending on whether the limits of the probabilities are changed.

The second way is provided by the CPT program, with the command TOOL/FORECAST/SERIES/ top part climate where the period assumed in the calculation is shown (Fig-ure 61).

Figure 63

C H A P T E R II

Period of the climatology

Low climatology level, in values.High climatology level, in values.

Low climatology level, in probabili-ties.High climatology level, in probabili-ties.

Example:Year of the forecast:2008Value of the forecast: 4.527ºCConfidence Interval: Low 2.980ºC High 6.075ºC



C H A P T E R II

This point is the number such that:

P[x ≥ Xα/2] – P[z ≥ Xα/2] – α/2

And in the standardized version:

Z-α/2= -Zα/2

Así:

Performing operations it is possible to clear μ to get the range:

[ ]P = 1 – α–Zα/2 ≤x – μ

σ√ n

[ ]P = 1 – αx–Zα/2 ≤ μ ≤ x + Zα/2σ σ

√ n √ n

Figure 64

Result is the confidence interval:

If σ is not known and n is large (i.e. ≥ 30):

s is the standard deviation of a sample.

Approximations for the value Zα/2 for standard confidence levels are 1.96 for 1 − α = 95% and 2,576 for 1 − α = 99%.

7. Where can I change the confidence level of my fore-casts?

Once the CPT is running, proceed to the following route: (Figure 64)

CUSTOMIZE/FORECAST SETTING/

( )x–Zα/2 , x + Zα/2σ σ

√ n √ n

( )x–Zα/2 , x + Zα/2s s

√ n √ n

8. How does the CPT consider probabilistic outcome 50% -10% -40% or 50% -0% -50%?

It is an ambiguity in which any of the scenarios is both pos-sible and not feasible, so it will only be regarded as uncer-tain. The CPT considers it with the average or normal value (normal category), but physically is not acceptable.

9. When do we consider a forecast with uncertain re-sults?Uncertainty is an expression of lack of knowledge of a fu-ture condition.

It can result from an absence of information or even because there is disagreement on what is known or what could be known. It can have multiple types of sources, from quantifi-able errors in the data to ambiguously defined terminology or uncertain projections of interpretation. Uncertainty can therefore be represented by quantitative measures (i.e., a range of values calculated by various models) or by quali-tative statements (i.e. reflecting the opinion of a group of experts). Within the CPT all results that have an obtained value correspond to the forecaster’s discretion.

10. How does the CPT consider a probabilistic outcome of 30%-40%-30%?

As explained in the previous question, these values are considered as uncertainties, i.e., any of the categories or conditions can occur under these conditions.

11. What is the cause of obtaining results with uncer-tainty?

It could be many causes, including:

Bad decision-making in the predictors used; which physi-cally explains the variability on the prediction (value to be predicted).

The CPT is based on the premise of the existence of a lin-ear relationship between predictors and predicting, which doesn’t always exist, which can be a cause of uncertainty.

The predictors are not defined because they are in a phase of changing astronomical station.

The poor quality of the information. In many cases the infor-mation from weather stations has fractures in the historical series because of significant changes in their location. Sta-tistically speaking this means that we have practically two different series that have been grouped for the run process with the CPT. (Figure 65)

Figure 65

The data series have many gaps; missing data also plays an important role in generating forecasts. The CPT program replaces the missing data values by mean values, medians, nearest station and at random.

Red Line: 10 followed years off


Figure 66

The modes are not adequate. Each mode carries a part of the variance to be explained from primary data (self values). Sometimes the number of adequate modes is not sufficient (usually within the top 5 modes is the explanation of a large percentage of the total variance). However, sometimes it is necessary to increase the number of modes to a recom-mended number of 10 (optional), with which the results are improved.

Rainfall in the countries near the equator is influenced by several simultaneous changes affecting precipitation and temperature variables. This requires working simultane-ously with multiple predictors (or different areas with only a single predictor).

12. How do we consider two opposite results obtained from two different predictive variables?

First verify if both have high CCC, and if they statistically acceptable; if both are correct, it is advisable to make an assembly with the predictors jointly, which we will have a result containing the two involved parameters in the vari-able to be predicted. Otherwise take the information from the higher CCC value.

13. How to perform simultaneous testing with two or more predictors?

The CPT is designed to take only one field of predictors at a time, but it is possible to obtain software in order to produce results with multiple fields. Run the software using one of the fields of predictors, and with the number of X EOF modes at maximum (this will be the minimum number of grid points and the length of test period). Then proceed to record the scores of principal components using Data Output. Repeat procedure for other predictor fields.

Then we proceed to combine multiple output files of the principal components scores so that the main components for all the predictor fields are in a file. CPT then can be run with this new file, as the variables read as a non-reference data set. Place in the X EOF the option “covariance matrix” to maintain the relative importance of EOFS. Although it will not be possible to see the maps for the merged fields, all validation results and forecasts will be as if the software would have been controlled with multiple input fields.

Some of the seasonal forecasts in the countries are shown in Figure 66.

C H A P T E R II



CHAPTER IIIImplementation of numerical models for climate prediction

CHAPTER III

Fig. 67 System Model MM51

1. University Corporation for Atmospheric Research, Weather Re-search and Forecasting Model users`s guide. Chapter 1http://www.mmm.ucar.edu//wrf/users/docs/user_guide_V3.1/users_guide_chap1.htm2. University Corporation for Atmospheric Researchhttp://www.mmm.ucar.edu//wrf/users/docs/user_guide_V3.1/users_guide_chap1.htm#WRF_Modeling_System

Fig. 68 WRF2 System

3.1.2.1 MM5

1. Download and install Intel Fortran

www.intel.comNote: There is a free non-commercial license. It is relatively

common to request a library libstdc + +. It is necessary to download (for example from pbone.net) and install it with a simple rpm.

2. Download and install NCAR

www.ucar.edu

Installation is simple. It is required to follow the Setup in-structions.

Note: It is suggested that you install it in: /usr/local/ncarg.

3. Download MM5

The packages needed are: TERRAIN, REGRID, LITTLE_R, INTERPF, MM5. ftp://ftp.ucar.edu/mesouser/MM5V3

4. Edit the /etc/bashrc

The last lines should say:

export PATH=$PATH:/opt/intel/fc/9.1.036/bin:/usr/local/ncarg/binexportLD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/fc/9.1.036/lib:/usr/local/ncarg/libexport NCARG_RO OT=/usr/local/ncarg

The situation showed above corresponds to an example. It is necessary to set the paths to the correct directories of the compiler. To load the newly introduced environment variables, it is enough to write: source/etc/bashrc.

5. To verify that the process is correct, consider the fol-lowing steps:

5.1. IFC: write ifort-v (It should display the installed ver-sion).5.2. NCAR: idt (It should open a graphical window)

6. Create a directory (e.g. /datos/MM5) and decom-press TERRAIN:

3.1 STEP BY STEP PROCEDURES FOR INSTAL-LATION AND IMPLEMENTATION OF MM5 AND WRF MODELS IN CLIMATE MODE

3.1.1 Operating System

The installation process (with images step-by-step) and the implementation of Scientific Linux, Rocks Cluster and Con-figuration and installation of a computer node is available at:

Scientific Linux: http://mediawiki.cmc.org.ve/index.php/imagen:Scilinux00.png

Rocks cluster and computer nod: http://mediawiki.cmc.org.ve/index.php/%E2%97%A6_Rocks_Cluster

3.1.2 Atmospheric Models

The atmospheric models considered in the Project are the fifth generation of the Mesoscale Model (MM5) and the Weather and Research Forecast model (WRF). The follow-ing pages show their installation and configuration. The same models, with appropriate modifications, are set as climate versions. These versions have been called CMM5 and CWRF.

The MM5 model is divided into multiple modules and sub-programs. Figure No. 66 presents a schematic diagram of MM5. The same way, in Figure No. 67 presents a diagram of the WRF model.

Angel Muñ[email protected]



C H A P T E R III

> cd /datos> mkdir MM5

> tar -xvzf TERRAIN.TAR.gz (clearly this file MUST be in

this directory)

7. Verify if the libg2c library is installed

If the libg2c library is not installed, proceed to install it now. If it has a different name, create the symbolic link.

Note: This library can be downloaded online, it is also avail-able in the gfortran. For instance, in [email protected]:

Another way: you can download it from: http://www.cmc.org.ve/descargas/libg2c.so

[root@Aquila TERRAIN]# find /usr ¬name “*libg2c*” This search the library/usr/local/matlab/sys/os/glnx86/libg2c.so.0/usr/local/matlab/sys/os/glnx86/libg2c.so.0.0.0/usr/lib/libg2c.so.0/usr/lib/gcc/i386¬redhat¬linux/3.4.3/libg2c.so/usr/lib/gcc/i386¬redhat¬linux/3.4.3/libg2c.a/usr/lib/libg2c.a/usr/lib/libg2c.so.0.0.0[root@Aquila TERRAIN]# ln -¬s /usr/lib/gcc/i386¬redhat¬linux/3.4.3/libg2c.so /usr/lib/libg2c.so

Place it in /usr/lib and perform an additional symbolic link as follows:

> ln -s /usr/lib/libg2c.so /usr/lib/libg2c.so.0

8. Edit the TERRAIN Makefile

Find the line that corresponds to the intel compiler and modify the PATH to lg2c:

> vi Makefile> /intel This finds the appearance of the word after the

slash.

The paragraph should be as follows:

intel:echo “Compiling for Linux using INTEL compiler”

( $(CD) src ; $(MAKE) all \“RM = $(RM)” “RM_LIST = $(RM_LIST)” \“LN = $(LN)” “MACH = SGI” \“MAKE = $(MAKE)” “CPP = /lib/cpp” \“CPPFLAGS = -I. C traditional D$(NCARGRAPHICS) “ \“FC = ifort “ “FCFLAGS = -I. -w90-w95-convert big_endian “\“LDOPTIONS = -i_dynamic” “CFLAGS = -I. “\“LOCAL_LIBRARIES=-L$(NCARG_ROOT)/lib -L/usr/X11R6/lib -lncarg -lncarg_gks-lncarg_c-lX11-L/usr/lib -lg2c” ) ; \( $(RM) terrain.exe ; $(LN) src/terrain.exe.) ;

9. Now we proceed to compile:

> make intel> make terrain.deck

10. Download the necessary data for TERRAIN as fol-lows and decompress it

> cd /datos/MM5/DATOS> wget ftp://ftp.ucar.edu/mesouser/MM5V3/TERRAIN_DATA/*> ls-1 > gunzip *.gz> tar-xvf archivo.TAR

10.1. Modify terrain.deck.intel

> vi terrain.deck.intel

And modify:

> set ftpdata = false> Set the following for ftp ’ in g30 sec> elevation data from USGS ftp site> set Where30sTer = /mnt/data/terrain_data

The result should be as follows:

#set ftpdata =trueset ftpdata = false#set Where30sTer = ftpset Where30sTer = /datos/MM5data/DATOS

Then proceed to link:

> ln -s /datos/MM5data/DATOS/* TERRAIN/Data/

11. Compile TERRAIN again and run

> make terrain.deck> ./terrain.deck.intel

Note: This compiles the code again. When finished, log into terrain.print.out and make sure the two last lines show:

> tail 2 terrain.print.out

If the process is correct, at the end of the run this should appear:

== NORMAL TERMINATION OF TERRAIN PROGRAM ==99999

Then write

idt TER.PLT

12. Create a download folder for “TERRAIN DATA”

Download from there as needed.

$cd $LOQUESEA/mm5$mkdir DATOSls$ cd DATOS$ wget ftp://ftp.ucar.edu/mesouser/MM5V3/TERRAIN_DATA/*


$ for x in ‘ls 1 *.gz‘; do gunzip $x; done

13. Unzip REGRID

In /datos/DatAquila/Meteo/mm5 and compile

$ make intel

Then download the data: NCEP_ON84.9303 in /datos/Meteo/DatAquila/mm5/DATOS, which is an input file for pregrid.

wget c –passiveftp ftp://ftp.ucar.edu/mesouser/MM5V3/TESTDATA/NCEP_ON84.9303

14. Log into pregrid folder

Edit pregrid.csh the lines that follow

set DataDir =/datos/DatAquila/Meteo/mm5/DATOS

15. Run pregrid.csh

$ ./pregrid.csh

It should read: **********Normal termination of program PREGRID_ON84**********mv SNOW:19930313_00 ../ON84_SNOW:19930313_00mv SNOW:19930313_12 ../ON84_SNOW:19930313_12mv SNOW:19930314_00 ../ON84_SNOW:19930314_00

Now cd /datos/DatAquila/Meteo/mm5/REGRID/pregrid/on84/..

If the process is right, this should appear in the pregrid di-rectory (the result of ls-l):

Doc/ nise/ ON84_SNOW:19930313_00 pregrid.csh*era/ nnrp/ ON84_SNOW:19930313_12 pregrid_era40_int.csh*grib.misc/ on84/ ON84_SNOW:19930314_00 pregrid.namelistMakefile* ON84:19930313_00 ON84_SST:19930313_00 RE-ADME_ERA40navysst/ ON84:19930313_12 ON84_SST:19930313_12 toga/ncep.grib/ ON84:19930314_00 ON84_SST:19930314_00 util/

16. Find in the pregrid directory the useful directory; there should be a file called plotfmt.

To compile the file, you should make the following changes to the Makefile:

NCARG_LIBS= ?L$ (NCARG_ROOT) /lib \ ?lncarg ?lncarg_gks ?lncarg_c \?L/usr/X11R6/lib ?lX11 ?lm \?L/opt/intel/fc/9.1.036/lib ?L/usr/lib ?lg2c

Then

$ make plotfmt

If there are no errors: $./plotfmt ../ON84:1993¬03¬14_00$ idt gmeta

Go to the regridder directory and:

$ ./regridder

17. If everything is correct, in the last step the following file will be created: REGRID_DOMAIN1

18. For LITTLE_R first proceed to decompress it

$ tar xvzf LITTLE_R.TAR.gz

(The file created should be placed in /datos/DatAquila/Meteo/mm5)

19. Log into the Makefile of LITTLE_R (in the intel op-tions)

-L/usr/lib/gcclib/i386redhatlinux/3.3.2

To -L/usr/lib lg2c.

It should read as follows: “LOCAL_LIBRARIES= -L$(NCARG_ROOT)/lib -L/usr/X11R6/lib -lncarg -lncarg_gks -lncarg_c -lX11 -L/usr/lib - lg2c” >> macros_little_r ; \ ( $(CD) src ; $(MAKE) $(PROGS) ) Note: Where the above was now reads -L/usr/lib -lg2c.

20. Download test data for LITTLE_R.

wget c –passiveftp ftp://ftp.ucar.edu/mesouser/MM5V3/TESTDATA/input2little_r.tar

Proceed to place it in: /datos/DatAquila/Meteo/mm5/DATOS then decompress the files as follows:

$ tar xvf input2little_r.tar

and the following files should be obtained: (ls-l)

Test_dataTest_data/REGRID_DOMAIN1.gzTest_data/surface_obs_r:19930313_21.gzTest_data/obs13_00.gzTest_data/obs14_00.gzTest_data/obs13_06.gzTest_data/surface_obs_r:19930313_18.gzTest_data/surface_obs_r:19930313_15.gzTest_data/surface_obs_r:19930313_12.gzTest_data/obs13_18.gzTest_data/obs13_12.gzTest_data/surface_obs_r:19930313_09.gzTest_data/surface_obs_r:19930313_06.gzTest_data/surface_obs_r:19930313_00.gz

C H A P T E R III



C H A P T E R III

Test_data/surface_obs_r:19930314_00.gzTest_data/surface_obs_r:19930313_03.gz

Place the data that you created in the folder TEST_ and write:

$ gunzip *.gz

It will get the following files:

obs13_00 obs14_00 obs13_12 obs13_06 obs13_18REGRID_DOMAIN1 surface_obs_r:19930313_06 surface_obs_r:19930313_18 surface_obs_r:19930313_09 surface_obs_r:19930313_21surface_obs_r:19930313_00 surface_obs_r:19930313_12 surface_obs_r:19930314_00surface_obs_r:19930313_03 surface_obs_r:19930313_15

All these files will be located in: /datos/DatAquila/Meteo/mm5/DATOS/Test_data

21. Modify namelist.input

The result should be:

&record2 fg_filename = ‘../REGRID/regridder/REGRID_DOMAIN1’ obs_filename= ‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/obs13_00’ ‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/obs13_12’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/obs14_00’sfc_obs_filename= ‘/datos/DatAquila/Meteo/mm5/DA-TOS/Test_data/surface_obs_r:19930313_00’ ‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930313_03’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930313_06’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930313_09’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930313_12’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930313_15’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930313_18’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930313_21’‘/datos/DatAquila/Meteo/mm5/DATOS/Test_data/sur-face_obs_r:19930314_00’ /

22. Run the test:

$ ./little_r

After a few minutes a pair of files will be created. In particu-lar, LITTLE_R_DOMAIN1 is necessary to run MM5.

Note: for each data it is necessary to modify namelist.input.

23. OPTIONAL: Install RAWINS

The installation will not be explained in this guide.

24. Install INTERPF

INTERPF is responsible for doing pressure interpolations.

Go to the MM5 directory and write (in this case the tar.gz is in the directory immediately above).

$ tar xvzf ../INTERPF.TAR.gz

25. Now simply

$ cd INTERPF$ make intel$ ./interpf

The above should create the files: MMINPUT_DOMAIN1, LOWBDY_DOMAIN1 y BDYOUT_DOMAIN1That will be used by MM5.

26. MM5:

It should begin by decontaining and decompressing: Go to the MM5 directory and

$ tar xvzf ../MM5.TAR.gz

Now go to the Run directory (which is within the MM5) and perform the following symbolic links:

$ ln s ../../INTERPF/MMINPUT_DOMAIN1 .$ ln s ../../INTERPF/BDYOUT_DOMAIN1 .$ ln s ../../INTERPF/LOWBDY_DOMAIN1 .$ ln s ../../TERRAIN/TERRAIN_DOMAIN2 .

27. Go back to the MM5 directory

And edit the section corresponding to 3I2 (INTEL with Intel Fortran Compiler) of configure.user.

The result should be shown as follows:

## 3i2. PC_INTEL (LINUX/INTEL)#RUNTIME_SYSTEM = “linux”FC = ifortFCFLAGS = I$(LIBINCLUDE) O2 tp p6 pc 32 convert big_endianCPP = /lib/cppCFLAGS = OCPPFLAGS = I$(LIBINCLUDE)LDOPTIONS = O2 tp p6 pc 32 convert big_endianLOCAL_LIBRARIES =MAKE = make i R


C H A P T E R III

28. Compile and run:

$ make$ make mm5.deck$ ./mm5.deck

If the process is successful this should be displayed:

Make [1]: Leaving directory `/datos/DatAquila/Me-teo/mm5/MM5/Run’This version of mm5.deck stops after creating namel-ist file mmlif.Please run code manually.vie mar 30 17:39:26 VET 2007

Now:

$ cd Run$ ./mm5.exe

1.- Settings:

In configure.user (/datos/CMM5/MM5/configure.user) all the information about the settings can be found (Section 6 of the file). In Section 5 of that file the parameters have to be considered carefully:

MAXNES = N (Here, the maximum number of domains to run in mm5.exe should be set).

MIX,MJX is the pre-dimensioning that is done for the ar-rays along the north-south and east-west axes. If a domain has been created in which the dimensions north-south or east-west exceed these parameters, you must increase MIX and MJX.

IMPORTANT: each time you change the configure.user should type: make clean; make (for the changes to take effect).

2.- On the other hand there is the mm5.deck (/datos/CMM5/MM5/mm5.deck).

Most important aspects to consider:

TIMAX = NNN (Total number of minutes that the forecast will last: NNN minutes to the future).

TISTEP = (is the delta T, in seconds the temporal integra-tion step. If CFL violations occur, this step should be decreased, and is linked to the spatial resolution chosen. The recommendation is to use a little less than 3 times the distance between the points assumed in TERRAIN for the thicker domain-the one with lower resolution).

3.- Other important options:

RADFRQ = 30. (Indicates how often atmospheric radiation subroutines are calculated in minutes. This value is appro-priate in order to start).

LEVSLP = 9, ;nest level (correspond to LEVIDN) at which solar radiation needs to be taken into account for orogra-phy; set the large to switch off; only have an effect for very high resolution model domains.

OROSHAW = 0, ;include effect of orography shadowing; ONLY has an effect if LEVSLP is also set; 0=no effect (de-fault); 1=orography shadowing taken into account - NOT AVAILABLE FOR MPI RUNS.

IMOIAV = 1, 1, Schematic of variable humidity. Depend-ing on the case, for weather, select 1 or 2; 0 - not used; 1 - is used without additional data; 2 - used with data from additional moisture.

OROSHAW controls whether or not to include shadow ef-fects due to orography in executions. Obviously, it is more physical, and costs more. If you wish to activate it , set LEVSLP, indicating the nest (1 = father, 2 = child, 3 = grand-son, etc.) from which OROSHAW begins to be used.

4.- Initial conditions:

An important aspect corresponds to the way of the analy-sis data is assimilated for the initial conditions. This is per-formed as follows:

IBOUDY = 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, ;boundary conditions; (fixed, time-dependent, relaxation -0,2,3)

IIf the domain is too large (all of Brazil, all of South America, etc ...) you should use a relaxed scheme of the boundary conditions (for references see the Online MM5 Manual or refer to Davies & Turner, Quart. J. Roy. Meteor. Soc, 103, 225-245 (1977)). For the remainder ones the time-depen-dent scheme can be used.

5.- The TSM item variable throughout the execution is also important.

It should be turned it in on in the following option: ISSTVAR= 1,

6.- This might be useful:

IFSNOW = 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ;SNOW COVER EFFECTS - 0, 1, 2; ;0 - no effect, 1 - with effect, 2 - simple snow model

7.- Now proceed to seek this section:

NEST AND MOVING NEST OPTIONS

LEVIDN = 0,1,2,1,1,1,1,1,1,1, “NEST-ED” LEVELNUMNC = 1,1,1,3,1,1,1,1,1,1, IDENT. MOTHER DOMAINNESTIX = 39, 13, 19, 46, 46, 46, 46, 46, 46, 46,NORTH-SOUTH SIZENESTJX = 45, 22, 13, 61, 61, 61, 61, 61, 61, 61,EAST-WEST SIZENESTI = 1, 20, 18, 1, 1, 1, 1, 1, 1, 1, ORIGIN IN NORTH-SOUTHNESTJ = 1, 13, 9, 1, 1, 1, 1, 1, 1, 1, ORIGIN IN EAST-WESTXSTNES = 0., 0.,900., 0., 0., 0., 0., 0., 0., 0., MINUTE THIS DOMAIN IS INITIALIZEDXENNES =259920.,259920.,1440.,720.,720.,720.,720.,720.,720., MINUTE THE CORRESPONDING EXECUTION ENDS.



C H A P T E R III

It is necessary to proceed to adjust each requirement that is requested, in accordance with the provisions in terrain.namelist

And just below, put the options as follows:

IOVERW = 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, ; overwrite nest input; 0=interpolate from coarse mesh (for nest domains);; 1=read in domain initial conditions; 2=read in nest terrain file

3.1.2.2 CMM5

1.- TERRAIN

Check that terrain. namelist registers as “NSTTYP” 1 for the first domain and 2 for others who are using it. This ensures two-way feedback in the mesh.

2.- PREGRID (inside REGRID): /datos/CMM5/REGRID/pregrid/pregrid.csh

The first is “decontain” (spread) the files to work on. For example,

tar -xvf archivo.pgb.f00.tartar -xvf archivo.grb2d.tartar -xvf A#####tar -xvf A#####

After this step, the following changes occur (this illustrates just one example, it should be adjusted according to the needs of the users):

set DataDir = /datos/2005/1ero

Here is PATH where the data is.

set InFiles = ( ${DataDir}/pgb.f00####*)

Instead of ### place the beginning of the numbers of the year in question. Ex: pgb.f000506*.

set SRC3D = GRIB # Many GRIB-format datasetsset SRCSST = $SRC3D

set InSST = (${DataDir}/grb2d0506*)

As before. Indicate the beginning of the files to be used. The * takes all related.

In this section, adjust the dates:

START_YEAR = 2005 # Year (Four digits)START_MONTH = 06 # Month ( 01 - 12 )START_DAY = 01 # Day ( 01 - 31 )START_HOUR = 06 # Hour ( 00 - 23 )

Note: It should begin in 06

END_YEAR = 2005 # Year (Four digits)END_MONTH = 06 # Month ( 01 - 12 )END_DAY = 30 # Day ( 01 - 31 )END_HOUR = 18 # Hour ( 00 - 23 )

Define the time interval to process.INTERVAL = 21600 # Time interval (seconds) to pro-cess.# This is most sanely the same as the time interval for# which the analyses were archived, but you can re-ally# set this to just about anything, and pregrid will

The (INTERVAL) step usually takes 6 hours. You can check the directory listing directly.

Finally:

set VT3D = ( grib.misc/Vtable.NNRP3D )set VTSST = ( grib.misc/Vtable.NNRPSST )set VTSNOW = ( grib.misc/Vtable.xxxxSNOW )set VTSOIL = ( grib.misc/Vtable.xxxxSOIL )

3.- REGRIDDER (inside REGRID): /datos/CMM5/RE-GRID/regridder/namelist.input

As previously discussed, if you carried out the exercises de-scribed, you should proceed to set dates in a basic way. And the ptop_in_Pa, which must match with the first_guess. If the installation process was followed correctly, the pro-gram should work without changes.

REMEMBER: Regridder is run once per domain

4.- INTERPF (In /datos/CMM5/INTERPF/namelist.input):

The first two lines must show the following:

&record0 input_file= ‘../REGRID/regridder/REGRID_DOMAIN1’ /

Here, later you vary the domains, an interpf run for each.

The following section may vary for some cases.

&record3 p0 = 1.e5 ! base state sea-level pres (Pa)tlp = 50. ! base state lapse rate d(T)/d(ln P)ts0 = 275. ! base state sea-level temp (K)tiso = 0./ ! base state isothermal strato-spheric temp (K)

This corresponds to the definition of the base state from which MM5 defines number of other variables/parameters. Detailed explanation with the equations can be found at: www.mmm.ucar.edu/mm5/documents/MM5_tut_Web_notes/INTERPF/interpf.htm

3.1.2.3 WRF

1.- Downloads:

www.mmm.ucar.edu/wrf/src/WRFV2.2.1.TAR.gz (WRF as such).www.mmm.ucar.edu/wrf/src/WPSV2.2.1.TAR.gz (WPS, the preprocessor).

Topography data: www.mmm.ucar.edu/wrf/src/wps_files/geog.tar.gz


C H A P T E R III

Additional necessary libraries: www.mmm.ucar.edu/wrf/src/wps_files/jasper-1.701.0.tar.gzwww.mmm.ucar.edu/wrf/src/wps_files/libpng-1.2.12.tar.gzwww.mmm.ucar.edu/wrf/src/wps_files/zlib-1.2.3.tar.gz

2.- These files should be stored, for example, in a folder called tars in /. Start decompression. The first things are extra in this case:

ZLIB:-----cd /opttar -xvzf /TARS/zlib-1.2.3.tar.gzcd zlib-1.2.3./configuremakemake install

JASPER:-------cd /opttar -xvzf /TARS/jasper-1.701.0.tar.gzcd jasper-1.701.0./configuremakemake install

LIBPNG:-------cd /opttar -xvzf /TARS/libpng-1.2.12.tar.gzcd libpng-1.2.12./configuremakemake install

3.- Now proceed with netcdf. To avoid confusion with the versions, we suggest downloading the version available on the server: www.cmc.org.ve/descargas/netcdf.tar.gz

3.1.- Place it in /TARS (or wherever you are placing tar con-tainers). Unzip:

tar -xvzf netcdf.tar.gzcd netcdf-3.6.2export FC=ifort./configuremake; make install

If the tar does not work with netcdf.tar.gz, write netcdf.tar.Z

4.- Check that the last step is correct, it is crucial for WRF. An ls /usr/local/include/netcdf* debe mostrar:

/usr/local/include/netcdfcpp.h/usr/local/include/netcdf.inc/usr/local/include/netcdf.h /usr/local/include/netcdf.mod/usr/local/include/netcdf.hh

Perform:ls /usr/local/lib/libnetcdf*

It should show:

/usr/local/lib/libnetcdf.a /usr/local/lib/libnetcdf_c++.la/usr/local/lib/libnetcdf_c++.a /usr/local/lib/libnetcdf.la

5.- Proceed now to /datos and create the CWRF folder and unzip it:

mkdir CWRFcd CWRFtar -xvzf /TARS/WRFV2.2.1.TAR.gztar -xvzf /TARS/WPSV2.2.1.TAR.gz

6.- Locate the WRFV2 folder

cd WRFV2

It is needed to add new lines to /etc/bashrc. These are described below:

export JASPERLIB=/opt/jasper-1.701.0export JASPERINC=/opt/jasper-1.701.0ulimit -s unlimited

To update the environment variables, proceed as usual:

source /etc/bashrc

From now on, you may follow one of two options. The first is to set from zero WRF and the second is to download the configuration file. In the same directory you should write:

./Configure

The following appears: ** WARNING: No path to NETCDF and environment vari-able NETCDF not set.** would you like me to try to fix? [y]

Choose “y” and include the PATH: /usr/local/include/usr/local/lib

Every time that it asks. If the process is successful, a menu appears (at the beginning it indicates that it recognizes the paths to the JASPER library):

Please select from among the following supported platforms. 1. PC Linux i486 i586 i686, PGI compiler (Single-threaded, no nesting) 2. PC Linux i486 i586 i686, PGI compiler (single threaded, allows nesting using RSL without MPI) 3. PC Linux i486 i586 i686, PGI compiler SM-Parallel (OpenMP, no nesting) 4. PC Linux i486 i586 i686, PGI compiler SM-Par-allel (OpenMP, allows nesting using RSL without MPI) 5. PC Linux i486 i586 i686, PGI compiler DM-Parallel (RSL, MPICH, Allows nesting) 6. PC Linux i486 i586 i686, PGI compiler DM-Parallel (RSL_LITE, MPICH, Allows nesting) 7. AMD x86_64 Intel xeon i686 ia32 Xeon Linux, ifort compiler (single-threaded, no nesting) 8. AMD x86_64 Intel xeon i686 ia32 Xeon Linux, ifort compiler (single threaded, allows nesting us-ing RSL without MPI)



C H A P T E R III

9. AMD x86_64 Intel xeon i686 ia32 Xeon Linux, ifort compiler (OpenMP) 10. AMD x86_64 Intel xeon i686 ia32 Xeon Linux, ifort compiler SM-Parallel (OpenMP, allows nesting using RSL without MPI) 11. AMD x86_64 Intel xeon i686 ia32 Xeon Linux, ifort+icc compiler DM-Parallel (RSL, MPICH, allows nesting) 12. AMD x86_64 Intel xeon i686 ia32 Xeon Linux, ifort+gcc compiler DM-Parallel (RSL, MPICH, allows nesting) 13. PC Linux i486 i586 i686, g95 compiler (Single-threaded, no nesting) 14. PC Linux i486 i586 i686, g95 compiler DM-Parallel (RSL_LITE, MPICH, Allows nesting)Enter selection [1-14] : 10

The selection must be “10”. If you desire to test WRF, select 7 (makes it impossible to create nesting) or 8 (with nests).

7.- Compiling the WRF. Once the above steps are done, proceed with the compiling:

./compile em_real > log.log

WRF is legendary for having a very long compilation. Wait at least 40 minutes. If you wish verify the state of compila-tion, perform a vi log.log in the same directory. Some notifications of compilation can be seen directly in the di-rectory when you send it to do the job. These are important, particularly if there is an error, a screen will appear where we ordered the “compile em_real.

8.- Testing the WRF. If the process is successful, a way to do preliminary test is:

ls run

You should see some symlinks: nup.exe, ndown.exe and especially real.exe y wrf.exe. If they are highlighted in red, something in the process has failed. Then you proceed to perform an additional test: a short run of WRF.

To do this, you need to download some test data in WRF intermediate format, available at:http://www.mmm.ucar.edu/wrf/src/data/jan00_wps.tar.gz

Another method is to do it directly into the terminal:

cd test/em_realwget -c http://www.mmm.ucar.edu/wrf/src/data/jan00_wps.tar.gztar -xvzf jan00_wps.tar.gz

The following is the initialization of WRF for testing. If the process is correct, these commands will not generate er-rors:

cp namelist.input.jan00 namelist.input./real.exe

When finished, run the WRF itself.

./wrf.exe

The process will take a few minutes. The execution prog-

ress (24 hours total) can be seen on the screen. This ex-ample performs downscaling time mode with two domains. If at the end we can see the message:

COMPLETED SUCCESFULLY

Then it means the installation of the WRF is correct.

9.- WPS (WRF PREPROCESING SYSTEM) Proceed to com-pile the WPS.

cd /datos/CWRF/WPS

You will be able to distinguish in these executable some di-rectories similar to those used in the WRFV2 runs. First run:

./configure

After selecting the appropriate option, run:

./compile

The compilation is considerably shorter than the WRF.

10.- Domain Wizard. There is a multiplatform application (in Java) that can be used to perform WRF preprocessing. It is a kind of GUI for WPS. It can be downloaded at:http://wrfportal.org/domainwizard/WRFDomainWizard.zip

You should place it either in the WPS or WRF and proceed to decompress:

gunzip WRFDomainWizard.zip

Then proceed to run:

./run_DomainWizard

It is important to remember and know precisely where each file is. We recommend creating, on WRF directory level (which contains the WPS and WRFV2), a directory called Domain (Dominios), where you can place the various do-mains that are created.

Hereafter it will be possible to perform a forecast run in mode time with WRF.

3.1.2.4 CWRF

Conceptually, the climate mode configuration is similar to the CMM5.

1.- The first is to tell the WRF to update TSM throughout an execution. This will create even additional files that may be read on the way.

Go to the WRFV2 directory and edit the namelist.input. The lines to be modified in each record are (if they do not exist, you must create them)

&time_controlauxinput5_inname = “wrflowinp_d<domain>”,auxinput5_interval = 180,io_form_auxinput5 = 2

&physics


C H A P T E R III

sst_update = 1,

With this, the real.exe will write a wrflowinp_d## file type for each active domain (DO NOT change the <domain>) with TSM information. The interval is in minutes.

If the process is right, after running real.exe you will ob-serve:

wrfinput_d01wrfbdy_d01wrflowinp_d01

If there is only a single domain, those corresponding to other domains will appear.

2.- There is available documentation on the parameteriza-tions of the WRF by J. Dudhia3. The changes to the physics of WRF (and CWRF) set out are shown below:

REMEMBER: different settings can be placed between do-mains, but you should always check the compatibility be-tween them. In some items the same settings are placed for all. This is not necessarily correct, but may be considered as an option:

&physicsmp_physics (max_dom) micro physics options= 0, without micro physics= 1, Kessler scheme= 2, Lin et al. scheme= 3, WSM 3-class simple ice scheme= 4, WSM 5-class scheme= 5, Ferrier (new Eta) micro physics= 6, scheme for graupel WSM 6-class= 8, Thompson et al. scheme= 98, scheme (to disappear) of simple ice NCEP 3-class= 99, scheme (to disappear) NCEP 5-class

The following are valid if mp_physics =! 0, to maintain Qv > = 0, and adjust the other fields of humidity to be less than or equal to a determined critical value.

mp_zero_out= 0, ; without adjustment of any humidity field= 1, ; except for Qv, all the other arrangements of humidity will be nulled= 2, ; Qv >=0, Every other arrangements of hu-midity will be nulled at certain limit.

mp_zero_out_thresh= 1.e-8 ; Critical value, under the same, all the humidity arrangements, ; except Qv, will be nulled (kg/kg)

ra_lw_physics (max_dom) option longwave radiation= 0, without longwave radiation= 1, rrtm scheme = 3, CAM scheme (adjust levsiz, paerlev, cam_abs_dim1/2 below)

= 99, GFDL (Eta) scheme; adjust co2tf = 1

ra_sw_physics (max_dom)option longwave radiation= 0, without longwave radiation option= 1, Dudhia scheme= 2, Goddard short wave= 3, cam scheme also must set levsiz, paerlev, cam_abs_dim1/2 (see below)= 99, GFDL (Eta) longwave (semi-supported) also must use co2tf = 1 for ARW

radt (max_dom)= 30, ; minutes between radiation physics callsrecommend 1 min per km of dx (e.g. 10 for 10 km)

nrads (max_dom)= FOR NMM: number of fundamental timesteps between calls to shortwave radiation; the value is set in Registry.NMM but is overridden by namelist value; radt will be computed from this.

nradl (max_dom)= FOR NMM: number of fundamental timesteps between calls to longwave radiation; the value is set in Registry.NMM but is overridden by namelist value.

co2tfCO2 transmission function flag only for GFDL radia-tion= 0, read CO2 function data from pre-generated file= 1, generate CO2 functions internally in the fore-cast

ra_call_offsetradiation call offset= 0 (no offset), =-1 (old offset)

cam_abs_freq_s= 21600 CAM clearsky longwave absorption calcula-tion frequency (recommended minimum value to speed scheme up)levsiz= 59 for CAM radiation input ozone levels

paerlev = 29 for CAM radiation input aerosol levels

cam_abs_dim1 = 4 for CAM absorption save array

cam_abs_dim2 = e_vert for CAM 2nd absorption save arraysf_sfclay_physics (max_dom) surface-layer op-tion (old bl_sfclay_physics option)= 0, no surface-layer= 1, Monin-Obukhov scheme= 2, Monin-Obukhov (Janjic) scheme

3.1.3 Oceanographic Models

3.1.3.1 ROMS

The Agrif version of Rom is easy to use, and is not very de-manding in the amount of data to start the runs. Agrifer

3. J. Michalakes, J. Dudhia et al. The weather Research and fore-cast model: Software architecture and performance. 11th ECMWF workshop on the Use of High Performance Computing in Meteo-rology, Reading U.K., 2004



C H A P T E R III

version has the advantage of being easy to install. It is nec-essary to copy the files downloaded from Romstools page and to copy them to HDD. After running matlab from the /Roms/Romstools/Run folder, almost everything can be done with a significantly good level of feedback. As soon as the Rutgerts version is fully understood, some particular comparisons will be made.

First you must go to ../Roms/Romstools/Run and modify via terminal using the preferred text editor, the romstools_param.m file within it. The first thing to edit is:

ROMS_title = ‘Pacifico’; % It is recommended to place the name of the area to be studied for better control of the ROMS_config = ‘CMC’; %. runs. A name for the type of configuration. (Later there will be other files that will have the same ROMS_title and ROMS_config).

Then proceed to place the mesh dimensions of the area to be studied by placing the coordinates of the place: % Grid dimensions: % lonmin = -148; % Minimum longitude [degree east]lonmax = -75; % Maximum longitude [degree east]latmin = -10; % Minimum latitude [degree north]latmax = 10; % Maximum latitude [degree north]

The resolution of the grid in degrees:

% Grid resolution [degree] % dl = 1; %maximum is 1, minimum used by the CMC, dl=1/32;Number of vertical levels (must be the same in param.h) % N = 32;

Then:

% Minimum depth at the shore [m] (depends on the resolution, % rule of thumb: dl=1, hmin=300, dl=1/4, hmin=150, ...) % This affect the filtering since it works on grad(h)/h. % hmin = 300; % % Maximum depth at the shore [m] (to prevent the generation % of too big walls along the coast) % hmax_coast = 500; % Slope parameter (r=grad(h)/h) maximum value for topography smoothing %

rtarget = 0.02; %0.025;% GSHSS user defined coastline (see m_map) % XXX_f.mat Full resolution data % XXX_h.mat High resolu-tion data % XXX_i.mat Intermediate resolution data % XXX_l.mat Low resolution data % XXX_c.mat Crude resolution data % coastfileplot = ‘coastline_l.mat’;

coastfilemask = ‘coastline_l_mask.mat’;

Finally the last section to modify in order to meet the mini-mum requirements for a run is:

% 6 Temporal parameters (used for make_tides, make_NCEP, make_OGCM) Yorig = 2008; % reference time for vector time% in roms initial and forcing files % Ymin = 2008; % first forcing yearYmax = 2008; % last forcing yearMmin = 1; % first forcing month

Mmax = 2; % last forcing month% Dmin = 1; % Day of initializationHmin = 0; % Hour of initializationMin_min = 0; % Minute of initializationSmin = 0; % Second of initialization% SPIN_Long = 0; % SPIN-UP duration in Years

To execute the run, it is necessary to conduct the follow-ing analysis (not all are essential but it is recommended to perform all of them, to ensure greater accuracy in results) ROMS file names (grid, forcing, bulk, climatology, initial).

Then from Matlab:

>>start>>make_grid>>make_NCEP>>make_clim>>make_bry(in the case that make_NCEP it was not possible to do the forcing and the bulk) >>make_forcing>>make_bulk

you return to the terminal where you run the executable jobcomp

./jobcomp

and finally

./roms roms.in

If the process is correct, from matlab write

>>roms_gui

and through the menu the roms_avg.nc file opens in ROMSFILES

3.1.4 Displayers

3.1.4.1 GrADS

1.- Create a directory in /usr/local/GrADS

2.- Download

wget ftp://ftp.ucar.edu/mesouser/MM5V3/MM5toGrADS.TAR.gz

3.- Unzip tar -xvzf MM5toGrADS.TAR.gz

3.1.4.2 Vis5D

TOVIS5D

1.- $tar -xvzf tovis5d. $tar.gz This creates the TOVIS5D folder.

2.- Edit the Makefile as follows:

linux: cd src/ ; $(MAKE) target \


C H A P T E R III

“FC = ifort” \“FCFLAGS =free DLINUX I.convert big_ endian” \“CCFLAGS =g DLITTLE DUNDERSCORE c” \“LIBS =Vaxlib “$(RM) tovis5d ; $(LN) src/tovis5d .Then tovis5d.csh must show as follows:

if ( ! e $1 ) then echo “The file $1 does not exist” exit 1 endif tovis5d $1 >&! tovis5d.log

3.- Compile

$ make linux

The options for the prediction: !/bin/csh f set echo cat >! user.in << EOF &userin view_times=0.,3.,6.,9.,12.,15.,18.,21.,24.,27.,gracetime_in_seconds=300.,model_version = ‘mm5v3 output’,new_fields = ‘the’, discard_fields = ‘RAD’, ‘PP ‘, interp_2_height = .true., output_terrain = .false. / &end

4.- Go to /datos/MM5Vis to type

$tovis5d MMOUT_DOMAIN1

It should show =========================== normally ended ===========================

Vis5D

1.- The program can be downloaded from the following links: ftp://ftp.ssec.wisc.edu/pub/vis5d5.1/vis5d5.1.tar.Z ftp://ftp.ssec.wisc.edu/pub/vis5d/vis5ddata.tar.Z 2.- Create the folder to be installed in: /usr/local/vis5d $tar -xvzfvis5data.tar.Z

This is created:

EARTH.TOPO LAMPS.v5d OUTLSUPW OUTLUSAM SCHL.v5d vis5d5.1

$tar -xvzf vis5d5.1.tar.Z clone.tcl label.tcl lui5 movie2.tcl README spin.tcl trajcol.tcl contrib highwind.tcl Makefile movie.tcl README.ps src user-funcs convert import listfonts Mesa NOTICE PORTING trajcol2.tcl util wslice.tcl

then tovis5d.csh must show the text as follows:

if ( ! e $1 ) then echo “The file $1 does not exist”

3.- Then proceed to the compilation: $ make linuxx $ make linuxopengl

If the machine has NvidiaTMGraphics Card, place:

$ make linuxnvidia.

4.- Access the /datos/MM5Vis directory to type:

$vis5d vis5d.file

The models shown are running experimentally in the vari-ous countries, some of which are shown in Fig 69.

Figure 69



C H A P T E R III

3.2 IMPLEMENTATION OF NUMERICAL MODELS FOR CLIMATE PREDICTION

The Regional Group of Numerical Modeling

The NMSs of the Andean countries have given a written approval to officially belong to a Regional Numerical Mod-eling Group (RMG). This group was created in Guayaquil in June 2008, and it is at the moment under the Techni-cal Coordination of Prof. Angel G. Munoz (attached to the Scientific Modeling Center of The University of Zulia and CIIFEN research associate), and under the institutional co-ordination of CIIFEN.

The Group constitutes an efficient mechanism to consoli-date the technical capabilities of those who use models in NMSs and thus sustain and improve what is obtained throughout this regional project.

Annex II includes the GRM Reference Terms and letters of support signed by the 6 Directors of the Meteorological Services.

Additionally, a wiki was developed for the installation of the operating system, available at: http://www.cmc.org.ve/wiki/


CHAPTER IVImplementation of Agro

Climatic Risk maps

CHAPTER IV

4.1. DEFINITION OF RISK1

Risk is defined as the combination of the probability of oc-currence of an event and its negative consequences1. The factors that comprise it are the threat and vulnerability.

Thread is a phenomenon, substance, human activity or dangerous condition that can cause death, injury or other health impacts, as well as property damage, loss of live-lihoods and services, social and economic disruption or damage to the environment1. Threat is determined by the intensity and frequency.

Vulnerability the characteristics and circumstances of a community, system or asset that make it susceptible to the damaging effects of a threat1.

With the mentioned factors is obtained the following for-mula.

RISK= THREAD . VULNERABILITY2

The factors that make up the vulnerability are exposure, susceptibility and resilience, expressing their relationship in the formula.

VULNERABILITY= EXPOSURE . SUSCEPTIBILITY RESILIENCE 2

Exposure is the disadvantaged due to location, position or location of a subject, object or system at risk.

Susceptibility is the degree of inherent fragility of a sub-ject, object or system to counter a threat and receive a pos-sible impact due to the occurrence of an adverse event.

Resilience is the ability of a system, community or society exposed to a threat to resist, absorb, adapt and recover from the effects of timely and effective manner, including the preservation and restoration of its basic structures and functions.

4.2. CONCEPTUAL MATHEMATICAL MODEL OF AGRO-CLIMATIC RISK

For agriculture climate risk estimation, the following for-mula was used:

*Vulnerability = [Susceptibility / Resilience]. Exposure.

The threat is made up of the relation of three climatic pa-rameters: precipitation, maximum temperature and mini-

mum temperature in a seasonal period (three months) and is based on the output of the statistical model. These parameters are considered as external factors affecting the crop phenological development, adverse effects of in-creased intensity and frequency with which they produce floods, drought, frost and excessive heat events whose ef-fects are negative for most crops.

As internal vulnerability elements of directly proportional crops, we considered that the exposure and susceptibility of the crop is inversely proportional to its resilience. The exposure of the crop was determined considering the loca-tion and environmental conditions in which it grows, and that for this case were: climate agriculture floor, season, tex-ture, slope, soil retention capacity, areas prone to erosion, flooding, landslides, frost and other specific conditions in the pilot area to determine how much the crop is exposed to the climate threat.

On the other hand, crop resilience is determined by the degree of weakness in the face of adversity climate at dif-ferent stages of development; for example in the case of corn, high temperatures stop the growth of the plantation, during flowering it can suffer more damage because high temperatures increase the number of sterile plants and de-creases the number of kernels per cob , i.e. that the climate damage leads to reduced growth of crops per hectare and a reduction in their field.

As the last component and inversely proportional in the agriculture climate risk measurement is the ability to cope with adverse weather conditions, expressed in this study by management practices that farmers have to deal with environmental hazards; an example is the development of drainage and irrigation canals to offset deadly floods.

In conclusion, agriculture climate risk estimation is estab-lished by the relationship of probable climatic effects. This is determined by the parameter of precipitation and tem-perature on crops, whose vulnerability is represented by the susceptibility of the crop at different development cycles as well as the ability to cope with adversity represented by farmer’s management practices and its relationship along with the crop’s exposure. This is represented mainly by the soil grain size characteristics, the presence of the crop in areas of recurrent adverse events such as floods and frost

4.3. COMPONENTS AND AGRICULTURE CLI-MATE RISK VARIABLES

Agriculture climate risk components are borrowed from the general formula of risk calculation2, these components be-ing threat and vulnerability. They are in turn composed by exposure, susceptibility and the ability of the crop to face the threat. Each component is described as follows:

1. UNISDR, Terminology on Disaster Risk Reduction 2009 for the concepts of risk, vulnerability and threat.

2.Marti Ezpeleta, A., 1993. Cálculo del Riesgo de Adversidades Climáticas para los Cultivos: Los Cereales de Verano en Monte-negros. p.264Dpto. de Geografía y Ordenación del Territorio, Universidad Za-ragoza.

Pilar [email protected]

Nadia [email protected]

AGROCLIMATICRISK

CLIMATETHREAD

CROPSUSCEPTIBILITY EXPOSURE

=CROP

RESILENCE

CROP *VULNERABILITY



C H A P T E R IV

4.3.1. Thread

The most adverse weather threats to crops are extreme events or persistent rainfall and temperature, with which floods, drought and frost is associated.

To evaluate the threat, the following values were consid-ered: precipitation, maximum temperature and minimum temperature above and below normal, subject to grada-tions for threat assessment introduced by each.

These variables were assigned proportionate with the level of prediction, according with what is generated from statis-tical models for seasonal forecasting.

4.3.2. Vulnerability

As the general formula of vulnerability states2, we calculat-ed the components of vulnerability, i.e. exposure, suscepti-bility and resilience, in this way:

Exposure

FloodsTo evaluate the exposure we considered soil texture (to infer the water-retaining capacity), flood risk areas and al-titude.

Depending on the capacity of soil to retain water and con-sidering the texture as the central element related to this ability, the following values were assigned for different tex-tural types:

Table 1.- Evaluation of the climate threat. Project ATN/OC-10064-RG

THREAT SCENARIOS

> 50% of normal

50% above normal

40% above normal

30% above normal

Between 10 and 20% above normal

Normal

Between 10% and 20% below normal

30% below normal

40% below normal

50% below normal

Less than 50% of normal

5

4

3

2

1

0

1

2

3

4

5

VALUE

Table 2.- Evaluation of texture. Project ATN/OC-10064-RG

TEXTURE

Very Fine

Fine

Media

Moderately coarse

Coarse

5

4

3

2

1

VALUE

Table 3.- Evaluation of frequency of flooding. Project ATN/OC-10064-RG

Table 4.- Evaluation of altitudinal zones to floods. Project ATN/OC-10064-RG

The areas prone to floods were evaluated as follows:

The altitude is valued based on a level that divides the up-per zones from the lower zones in the area of interest, as follows:

FLOOD FREQUENCY

Very often

Often

Regularly

Little

Slightly

No

5

4

3

2

1

0

VALUE

ALTITUDE

High zone

Lower zone

1

2

VALUE

FrostTo evaluate the exposure, altitude and frost-prone areas were considered.

Frost-prone areas were evaluated as follows:

Table 5.- Evaluation of frequency of frost. Project ATN/OC-10064-RG

The altitude is valued based on a level that divides the zones of the area of interest, as follows:

FREQUENCY OF FROST

Very often

Often

Regularly

Little

Slightly

No

5

4

3

2

1

0

VALUE

Tabla 6.- Valoración de pisos altitudinales ante heladas. Proyecto ATN/OC-10064-RG

SusceptibilitySusceptibility was valued according to the phenological stage of the crop, for different possible climate conditions predominant development stage in which the crop is found for the month or period of interest. The assessment was performed considering the levels of precipitation and tem-peratures above and below normal.

ALTITUDE

High zone

Lower zone

2

1

VALUE


C H A P T E R IV

Table 11. Assessment of irrigation and drainage infrastruc-ture. Project ATN/OC-10064-RG

The evaluation was estimated considering the team of ag-ricultural experts of the project and the bibliography used [3].

ResilienceFor the assessment of resilience we considered the irriga-tion and drainage infrastructure, whose presence allows crops to reduce the impacts caused by adverse climate events; it is evaluated as follows:

4.4. PROJECT APPLICATION AREAS

For the definition of pilot areas the following criteria were considered:

• Existence of an acceptable spatial coverage of meteoro-logical stations.• Agricultural activity relevant in social and economic terms.• Farming activity with a level of vulnerability.• Available information base.

The pilot areas designated for the project by each of the countries with the selection of crops, are shown in the table below.

Table 7. Susceptibility rating of phenological phases to above normal rainfall. Project ATN/OC-10064-RG

PRECIPITATION

% Above normal Normal

Sowing-Germination

Maturation-Harvest

Growth-Tillering

Flowering

>50 50 40 30 20 10

Grain filling

4

5

5

5

5

4

4

5

5

5

3

4

4

5

5

2

3

4

4

5

1

2

3

4

5

1

1

2

3

4

0

0

0

0

0

STAGE

Table 8. Susceptibility rating of phenological stages to below normal rainfall. Project ATN/OC-10064-RG

PRECIPITATION

% Bellow normal Normal

Sowing-Germination

Maturation-Harvest

Growth-Tillering

Flowering

>50 50 40 30 20 10

Grain filling

5

5

5

5

4

5

5

5

4

4

5

5

4

4

3

5

4

4

3

2

4

4

3

2

1

4

3

2

1

1

0

0

0

0

0

STAGE

Table 10. Susceptibility rating of phenological stages at temperatures below normal. Project ATN/OC-10064-RG

TEMPERATURES

% Bellow normal Normal

Sowing-Germination

Maturation-Harvest

Growth-Tillering

Flowering

>50 50 40 30 20 10

Grain filling

5

5

5

4

3

5

5

5

3

3

4

4

4

3

2

4

3

3

2

2

3

3

3

1

1

3

2

2

1

1

0

0

0

0

0

STAGE

Table 9. Susceptibility rating of phenological stages at temperatures above normal. Project ATN/OC-10064-RG

TEMPERATURES

% Above normal Normal

Sowing-Germination

Maturation-Harvest

Growth-Tillering

Flowering

>50 50 40 30 20 10

Grain filling

4

5

5

5

4

4

5

5

5

4

3

4

4

4

3

2

3

4

3

2

1

3

3

3

1

1

2

3

2

1

0

0

0

0

0

STAGE

IRRIGATION AND DRAINAGE INFRASTRUCTURE

Presence

Abscence

1

2

VALUE



C H A P T E R IV

Table 12. Pilot Areas and Crops for Country. Project ATN/OC-10064-RG

Figure 70. Pilot Areas of Project ATN/OC-10064-RG.

Venezuela

Country

Portuguesa State

Pilot Area

Rice, corn, sesame, sorghum

Crops

Colombia Bogota and Tolima Flowers, rice

Ecuador Guayas, Manabí, Los Ríos Corn, rice, soybeans

Perú Mantaro Valley Potato, corn, artichoke

Bolivia Highland Region Potatoes, lima beans, quinoa

Chile Valparaíso Region Citrus, avocado


The forecasts of precipitation, maximum and minimum temperature, are generated by NMHSs and must be con-verted to digital format and georeferenced.

It requires for the satellite images of the area of interest to be updated, preferably within the last two years. The best image resolution required is10 m.

Table 17. Requirements of climate forecasting. Project ATN/OC-10064-RG

ClimateInformation

Maximum temperature forecast

Precipitation forecast

Minimum temperature forecast

Table 18. Requirement of satellite images. Project ATN/OC-10064-RG

SatelliteImages

Area ofinterest

4.5.4. Treatment of Information

The information collected must go through a validation process, correction, editing and standardization, which is known as information processing. In this process all the errors and discrepancies that exist on the provided infor-mation are corrected. It was necessary to standardize the layers to the same reference and projection system [WGS 84 - UTM].

4.5.5. Soil and climatic characteristics in pilot areas

As a result of data gathering, the edapho-climatic charac-terization was also obtain in each area.

4.5. INFORMATION REQUIREMENTS

Information was requested in tabular format for qualitative and quantitative information related mainly to agro-eco-logical characteristics of crops and in digital format [shape-file] for the base and thematic mapping.

4.5.1. Agro-ecological

The data required as inputs to evaluate the crops consider-ing their agro-ecological characteristics are:

• Predominant varieties.• Phenological stages.• Threshold precipitation [mm].• Threshold temperature [° C].• Annual Season.• Soil Texture.

We worked with the dominant variety that was the most representative of the crop in the area, and the precipita-tion and temperature thresholds that are required to find the optimal conditions for crop development. We included other requirements as the periods of the year for planting and harvesting and optimal soil texture. An example of agro-ecological requirements is shown in Table 14.

Table 14. Requirements and agro-ecological parameters. Project ATN/OC-10064-RG

Table 15. Requirements of basic cartography. Project ATN/OC-10064-RG

REQUIREMENTS AND AGRO-ECOLOGICAL PARAMETERS

Country

Pilot zone

Cultivation

Predominant varieties

Phenological stages

Threshold precipitation (mm) From: To:

Threshold temperature (ºC) From: To:

Economic threshold (%)

Annual season

Soil texture

4.5.2. Base mapping

Digitalized basic information was required to put together the base mapping. This information had to have official sta-tus and be as updated as possible.

4.5.3. Thematic mapping

Thematic information, that serves as input for this calcula-tion, should be obtained in digital format [shapefile] and be properly geo-referenced.

Table 16. Requirements for thematic mapping. Project ATN/OC-10064-RG

Thematic Mapping

Erosion

Floods

Landslides

Droughts

Frost

Current land use

Vegetal cover

Soil texture

Crop location

Basic Mapping

Provincial or Departmental Political Boundaries

National Political Limit

Municipal and Cantonal Political Boundaries

Water System

Road System

Populated Centers

Urban Areas

Contours

Topography

C H A P T E R IV



Table 19. Soil and climatic characteristics of the state of Portuguesa, Venezuela. Project ATN/OC-10064-RG

Venezuela

Colombia

Table 20. Soil and climatic characteristics for the Bogota Savannah and Tolima, Colombia. Project ATN/OC-10064-RG

Country Zone AltitudinalZones

Altitute(m.o.s.l.)

NormalPrecipitation(mm)

NormalMaximumTemperature (ºC)

NormalMinimumTemperature(ºC)

Soil Texture

Venezuela PortuguesaState

High Zone

Low Zone

<200

>200

1350

1250

31

32

22

22

Sandy, clay, silty clay, clay

Clay loam, silty clay, clay


Altitute(m.o.s.l.)




Soil Texture

Colombia

BogotaSavannah

Tolima

SouthwestSavannah

CenterSavannah

NorthSavannah

South Tolima

CenterTolima

SouthTolima

2543

2540

2580

425

431

450

220

200

200

550

525

500

22

22

22

35

34

34

8

8

8

22

20

20

Being flowers in gre-enhouses, for the Bo-gota area this parame-ter was not required

Clay loam

Sandy

Sandy clay loam

Table 21. Soil and climatic characteristics for Coast Region, Ecuador. Project ATN/OC-10064-RG* In the rainy season

Ecuador


Altitute(m.o.s.l.)




Soil Texture

Ecuador CoastRegion

UpperBasin

LowerBasin

>40

<40

*1500

1250

31

32

22

22

Clay loam, sandy loam

Sandy clay loam, silty loam, sandy loam

Table 22. Soil and climatic characteristics for the Mantaro Valley, Peru. Project ATN/OC-10064-RG* In the rainy season

Peru


Altitute(m.o.s.l.)




Soil Texture

Perú MantaroValley

UpperBasin

LowerBasin

>3350

<3350

*1100

*1000

19

20

5

6

Clay loam, sandy loam

Sandy clay loam, sil-ty loam, loam, sandy loam

C H A P T E R IV


Figure 71. Variables for a g r i c u l t u r e climate risk assessment . Project ATN/OC-10064-RG

CLIMATICDATA

(Threadevaluation)

ALTITUDINALFLOORS

(Level curves)SOIL TEXTURE,

EXPOSED ZONESTO FLOODS AND

FROSTS(Exposureevaluation)

FENOLOGICSTAGE

(Susceptibilityevaluation)

IRRIGATION AND DRAINAGE

INFRASTRUCTURE

(Resilienceevaluation)

ture periods (3 months). The first step was to summarize or simplify if necessary the description field of the attributes table of the corresponding variable. After this a new field was created, in where the values previously presented in Table 1 were assigned. Once they were evaluated, it was proceeded to simplify the attribute table for each variable, leaving only the fields of description and evaluation. To this point, the subsequent operations conducted with them were facilitated [union].

With the simplified attribute tables, we proceeded to join the three variables, obtaining a new layer of climatic threat conditions for the corresponding period, which summarizes in each polygon a homogeneous condition of precipita-tion, maximum temperature and minimum temperature (as can be illustrated in Figure 72).

Then, in the table resulting from the union of the three vari-ables, a new field is introduced, where the sum of the value of each variable[the three areas of evaluation] is made; this implies the threat level that each area has and thus the component threats are now ready.

Exposure While the parameters involved in the exposure to rain or extreme temperatures are more or less stable over time, an exposure map was prepared, which will be considered as a constant for the next few months.

In the case of texture, records were generalized (summa-rized) based on the field that describes the texture and then a new field was created for the assessment of each texture category, using the values in Table 2.

Table 23. Soil and climatic characteristics for Region Altiplano, Bolivia. Project ATN/OC-10064-RG* Annual Average

Bolivia


Altitute(m.o.s.l.)




Soil Texture

Bolivia PlateauRegion

NorthernHighlands

CenterHighlands

SouthernHighlands

4000

3500 to 4500

3500 to 4500

*660

*429.2

*247.8

10.7

11.9

16.8

6.8

5.7

7.9

Silty loam, clay loam

Sandy, sandy loam, sil-ty loam-clay

Sandy loam

Chile

The soil and climate parameters in Valparaiso were pro-vided in digital data layers (shapefile format). Agro-climatic zones in the Valparaiso Region belonging to semi-arid and temperate system are:

• Semiarid Andes Mountains• Semiarid Middle Mountain• Semiarid Northern Coastal• Temperate Andean Cordillera• Temperate Intermediate Depression• Temperate Coastal Range• Temperate Coastal Central• Temperate Southern Coast

In the Valparaíso Region there are four types of climate: a dry steppe climate which is the continuation of the climate in the IV Region and three temperate climates that are dis-tinguished from each other by the characteristics of clouds and the length of dry periods. Its average annual rainfall varies in its various zones between 260 and 560 mm. Tex-tured soils are mostly sandy-clay and silty-sandy.

4.6 AGRO-CLIMATIC RISK CALCULATION

For the agricultural climate risk evaluation, a GIS tool was used to calculate all its components, using the variables in-herent to each of them as illustrated in Figure 71.

Thread MapThe calculation of the climate thread starts with the rainfall forecast, maximum temperature and minimum tempera-

C H A P T E R IV



Figure 72. Polygons of climatic threat conditions. Project ATN/OC-10064-RG

In the case of flood risk, its attribute table records were generalized (summarized) based on the field that describes the risk of flooding and then a new field was created for the evaluation of each category of floods, using the values set in Table 3.

For the case of altitude, its records were generalized (sum-marized) based on the field that describes these soils and a new field was created for the values of each of its catego-ries, using the values set forth in Table 4.

The table of the attributes of each one of the variables list-ed above was simplified so as to show only those essential elements, or those fields related to both the description of each parameter as well as its value, and thereby simplify the process of “union” described next.

Finally, we proceeded to the union of these three layers and the resulting attribute table of this union, a new field of to-tal exposure evaluation, is created, which will be obtained through the sum of partial valuation fields of the three vari-ables (risk of flooding, texture, altitude levels) for each re-cord or polygon.

To simplify the Exposure map for subsequent processes, reduce the table leaving only the field of exposure evalu-ation (sum). The map will become the constant exposure for some time, due to the low temporal fluctuation of its variables.

Susceptibility For the evaluation of the susceptibility, we work with the union of the corresponding crop layer and the layer of ho-mogeneous climatic conditions. In the attribute table of the weather, a new field properly summarized [generalized] was introduced, which will give the values of crop susceptibil-ity to these climatic conditions. The valuation tables of the susceptibility are found in the tables: 7, 8, 9 and 10. We should repeat the same procedure for each crop, thus the component of susceptibility will be solved.

Capacity for Recovery / ResilienceTo obtain a resilience map, it was necessary to rely also on the crop layer, adding information about the presence or absence of irrigation or drainage canals. Records are generalized based on the field that describes the existence

or non-existence of infrastructure. Later a new field is cre-ated for the assessment of the two infrastructure categories [whether or not it exists], using the values set forth in Table 11.

In all cases after the evaluation, we must simplify the at-tribute tables so as to show only those essential elements, or those fields related to both, the description of each pa-rameter and its value, in order to simplify the calculations on which these variables intervene.

Vulnerability mapProceed to join the layers of susceptibility, resilience and exposure. In the attribute table of this union a field was added, where the processes established in the formula of vulnerability will be applied. For example, multiply the ex-posure field value by the susceptibility field value and di-vide it by the resilience field value.

To simplify the vulnerability map, to be used in the latest risk measurement process, the table is cut down, leaving only the last field where the formula for calculating vulner-ability was developed.

Agro-Climatic Risk MapThe next step was to link the vulnerability map with the threat map and to multiply the fields of implicit valuation in each of these 2 components, presenting these results on a map with the estimated resulting risk.

Figure 74. SIG structure for Agro-Climatic Risk calculation. Project ATN/OC-10064-RG

C H A P T E R IV

Table 24. Agriculture climate risk assessment. Project ATN/OC-10064-RG

RISK LEVEL

R G B

High

Moderately High

Medium

Moderately Low

Low

5 168 0 0

4 230 0 0

3 255 70 70

2 255 127 127

1 255 190 190

COLORVALUE

The agricultural climate risk level obtained can be repre-sented by its absolute values or by risk intervals. We rec-ommend assigning different shades of red to the different levels of risk.


4.7 AGRO-CLIMATIC RISK IN THE ANDEAN COUNTRIES

The agricultural climate risk assessment in the Andean countries is carried out with the application of the devel-oped methodology that allows the integration of basic vari-ables for risk calculation in each country. The methodology used generates the model for a first risk approximation, which, although is an estimate, it gives a tool to support decision making in the agricultural sector. Each participat-ing country in the project made adjustments to some of the proposed variables in order to obtain results tailored to local realities and therefore stating that this methodology gives us the base guidelines to obtain a first approximation of agriculture climate risk and it should have adequations and adjustments as required.

On next page are the agricultural climate risk maps created for the 6 countries.

C H A P T E R IV



Figure 75. Agriculture climate Risk Map of Sesame crop. Estate of Portuguesa, Venezuela 2008. Project ATN/OC-10064-RG

C H A P T E R IV


Figure 76. Agriculture climate Risk Map of rice crop. Tolima Valley, Colombia 2008. Project ATN/OC-10064-RG

C H A P T E R IV



Figure 77. Agriculture climate Risk Map of rice crop. Costa de Ecuador 2008. Project ATN/OC-10064-RG

C H A P T E R IV


Figure 78. Agriculture climate Risk Map of potato crop. Mantaro Valley, Peru 2008. Project ATN/OC-10064-RG

C H A P T E R IV



Figure 79. Agriculture climate Risk Map of potato crop. Plateau of Bolivia 2008. Project ATN/OC-10064-RG

C H A P T E R IV


Figure 80. Agriculture climate Risk Map of citrics crop. Valparaíso region, Chile 2008. Project ATN/OC-10064-RG

C H A P T E R IV



CHAPTER VImplementation of local

systems of climateinformation

CHAPTER V

5.1 CONCEPTUAL AND METHODOLOGICAL ELEMENTS

The confusion generated by the users for climate informa-tion creates distrust, even more in a sector so vulnerable to the climate as is farming in South America and the conside-rable climatic dependence for irrigation. The producers of climate information mistakenly assume that the information they provide will be absolutely crucial in making decisions by users, in this case farmers. Seen from the perspective of the users, the picture is different. The decision-making process is anthropologically determined by a set of infor-mation in which the climate is a part of, but it is not the only component. It needs to consider the atmosphere, the attitude and finally a totally unpredictable element that de-pends on the culture, perception and psychological profile of the users.

Although the decision maker is always going to look for more and more detailed weather information, the role of this information will be shared with other elements not related to climate. What is relevant in this analysis is that something that cannot happen is that the decision-making process does not use climate information at all because it does not have it, or it does not understand it, or because it is confusing or simply because he does not trust the infor-mation. In this context, from the standpoint of the mental processes involved, it is much more valuable to have mo-dest information in resolution, but clear and accessible, so that it is used in the process. The challenge is to ensure that the farmer will give climate information its space in his decision making-process. To achieve this would represent a basic pillar in integral management of climate information.

The premise of optimum weather information management comes from the fact that instead of having few informed people (usually scientists) with forecasts and good-quality climatic information, we should have informed people with acceptable weather information. This would mean more people making decisions based on the modest but accep-table weather information provided, but delivered in such a way that it is fully used1.

The implementation of local systems for the dissemination of climate information is intended to deliver this informa-tion to the farming community through various sources (print, radio, magazines, newsletters, television, sms text messages, email, among others) through the consolidation of user networks, strategic alliances, training workshops and capacity building to promote the system and especially the products. The climate services used on this project (da-tabase, climate, statistical and numerical forecasts, agro-climatic risk maps) for each country, plus the existing pro-ducts at each NWS should ensure timely, fast, reliable and over time sustainable distribution, in which the information is not distorted and it serves to support decision making.

1. Martinez, Rodney, 2006. Information management and climate prediction services to reduce impacts on agriculture in South Ame-rica. Campinas, 8-15.

2. Martinez, Rodney, 2006. Information management and climate prediction services to reduce impacts on agriculture in South Ame-rica. Campinas, 8-15.

The climate information dissemination systems are desig-ned to successfully close the cycle of information manage-ment. This envolves the design of strategies for sustainabi-lity and consolidation over time.

The strategy used to strengthen the climate information system includes the following lines of action:

1) Strengthen the final format of climate information pro-ducts.2) Articulate the means to disseminate the information.3) Empower users by introducing and involving them in the system, and4) Establish alliances with potential actors/beneficiaries sys-tem multipliers to strengthen it.

Figure 81 shows the cycle of information management. The conversion of the products to a simpler language and user-friendly and specific to each media format allows informa-tion to be distributed in many forms, and makes it more likely to be assimilated. These media can range from televi-sion, newspapers, radio, internet, and even text messages via cell phone or HF radio.

The way in which users in a country see the weather is not in a computer chart or in the output of a model; users see the climate as what they experience: rain, drought, frost, wind, etc. If this perception is later associated with a name, for example: El Niño, La Niña, a physical pattern is generated in the imagination of the user, which is experienced by a label. This is internalized and remains in the minds of the users. Now, after a while when referring to El Niño or La Niña, for the user, there are only images associated with floods or droughts, death and destruction; the rest of the text that is used to supplement the information is simply transparent to them: it does not exist, it is not assimilated, it is just the mental image of what was internalized, and they will act accordingly with it if the message is repetitive or convincing2. The disseminated information is effective when it is understandable; it is received without distortions and generates a response in the recipient, for this, there must be a network of key users to maximize its distribution.

The information can be distributed to different user groups, whether these are authorities, representatives of associa-tions or unions, rescue teams (fire department, Red Cross), disaster management agencies, private sector, researchers and students of higher educational institutions, community representatives, among others. Due to differences in the distinctions of user groups, for purposes of a more standar-dized management at the regional level, they are classified into three main groups or categories and focused on the ul-timate goal of this regional initiative: the agricultural sector.

This structure was applied to the six Andean countries, emphasizing in a greater or lower level its component ca-tegories according to the socio-cultural and political inter-vention in each project region.

Abigail [email protected]

Alexandra [email protected]



C H A P T E R V

3. The methodology outlined is based on the document: Tools to Support Participatory Urban Decision Making Process: Satakehol-ders Analysis. Urban Governance Toolkit. UN-HABITAT program, 2001.4. It was not considered the fourth area, recovery whose compo-nents are rehabilitation and reconstruction.

5.2 IDENTIFICATION AND MAPPING OF KEY ACTORS

The mapping of actors is a technique used to identify all persons and entities that may be important to build a distri-bution network of weather information. This technique en-sures that mapped users clearly know beforehand to who they have to define specific strategies to help them ensure the flow of information so that the actions taken are coor-dinated3.

To perform a basic stakeholder mapping, you must perform the following steps: define the issues, identify stakeholders and map the actors.

• Define the topic

At this stage we specify which are the persons, groups or organizations on whom we should work according to the topic. They become important players to the work that is going to be performed.

In this case, the climate services generated are focused on agro-climatic risk management in three of the four areas [4] that includes it: Risk Analysis, Risk Reduction and Manage-ment of Adverse Events.

• Identify the actors

To identify the actors, several activities were carried out:

LISTAt this early stage we must work together to review any in-formation gathered and then, through brainstorming, have a list of all the persons or institutions that can meet the fo-llowing characteristics:

• Are being or could be affected by the problem.• Could be affected by the proposed solution to the pro-blem presented by the group.• They are not being directly affected, but may have an in-terest in the proposal.• They have information, experience or resources necessary to contribute to the goals of the project.• They have a national or local reach, for example associa-tions of farmers, rescue groups, private companies of agri-cultural products, among others.• They are accepted by the community, e.g. community ra-dio stations or radio fans, community leaders.• They are necessary for the implementation of project ac-tivities.• They feel entitled to be involved.• They are necessary for project sustainability.

Thus, we obtain a preliminary list of stakeholder groups that should be mapped:

FOCUSThe next step is to have each one of the groups identified and to obtain their contact information.

CATEGORIZEOnce all the information required of the members of each group is obtained, we will proceed to organize it into ca

Figure 81. General structure graph Climate Information System

Figure 82. Risk Management, figure by Omar Dario Car-dona, Adapted by CIIFEN, 2009

Figure 83. Preliminary group of key actors

Figure 84. Detail of the key actor

RiskAnalysis

CommunitaryLeaders

Communica-tion Media

Private/Productive

Sector

Private Sector

Agripac Corp. Group

Agripac S.A.

Public Relations

Cynthia Baratau, Agripac En Directo magazine PublisherAddress: Córdova 623 and Padre Solano.Phone (593-4)2313327 E-mail: [email protected]

StateAgencies

andAuthorities

Universities Internatio-nal

Agencies

RiskReduction

Manage-ment of Adverse Events


C H A P T E R V

tegories which in turn will have more subcategories. In this case three major groups were established: Authorities, Media and Pro-ductive Sector, as outlined below:

Figure 85. Mapping cate-gories and subcategories of Actors

Figure 86. A c t i v i t i e s for the stakeholder mapping and monitoring

Each country updated the contact information available in the National Weather Service, classified it in the prescribed format and then fed the database with new key users accor-ding to available and selected climate products following the intervention area within the country. The monitoring of each group is important to create an alliance and show the commitment, scope and applicability of the project.

After completing this step, we should contact each one of them; this represents field work to validate and comple-ment the previously formed database. Figure 79 shows two steps to carry this out: through meetings and their subse-quent monitoring. The meeting with the mapped actors constitutes the first advance, for which printed and digital material should be brought along to briefly report the pur-pose of the visit, and to form alliances, emphasizing that the benefit is mutual.

5.3 STRATEGIC ALLIANCES

There are several things to consider when setting a strate-gic alliance or commitment to cooperation through a letter of intent or commitment letter:

• These alliances are not contracts.

• The letters of commitment or intent formalize the part-nership of cooperation between the agency and the NWS (in this case).

• Letters of commitment should have equal obligations of both parties (win-win).

In Annex III there is the inventory of strategic alliances in the region.

It is recommended to manage these letters or agreements during the mapping of key players because at this stage contact or dialogue is direct. Moreover, achieving a strate-gic alliance and formalizing by means of a letter of intent is a process that is usually not achieved in the short term. It is advisable to carry out these partnerships with entities that correspond to any of the three proposed groups; however, there are some exceptions, such as remote locations or po-pulations that do not receive all the radio frequencies and therefore, have very specific services of certain frequen-cies that cover only that area. These local radio stations, in the case of broadcasting information to very vulnerable populations, become strategic partners when issuing an early warning or broadcasting climate information that the people need. Continuing with the same example, in these cases local radio stations tend to relay information for a li-mited time from other frequencies through phone or HF ra-dio. In the case of signing an agreement with a radio station that has this type of retransmission mechanism through a signal relay to a local radio or amateur radio, the impact of providing climate services has a wider scope.

For the National Weather Service, this represents the res-ponsibility to comply with sending information continuously and as stated in the agreement, meeting deadlines, format, length, and even ensure that it is in an easy-to-understand language.

The commitment letters provide a mechanism to ensure dissemination of information; in the case of climate, to a certain group of end users. Each entity, whether it is gover-nmental, private or non-profit, is committed to publicizing

• Coordination of the Meeting.• Creation of material for the presentation.• Information exchange through contact.• Establishment of verbal commitment.• Give information to those present.

CONTACT MAP(identification of actors)1

MEETING FORACTOR MAPPING2

• Sending of letters of intention to formalize alliances.• Sending of weather information created by CIIFEN.• Coordination of dates and places for workshops

MONITORINGKEY ACTORS3

Authorities Productive SectorCommunicationMedia

• National Government• Ministries• Regional Intendences• Sub secretarials• Government ministry• Township• Mayors• Red Cross• Firemen Deps.

• Cell Phones• Television• Radio National Newspapers• Local Newspapers• Internet

• Associations• Fraternities• Private Bussiness Production Chambers• Corporations• Universities• Communities



C H A P T E R V

Figure 87. Monitor ing of key pla-yers

• Sending of letters of intention to formalize alliances.• Sending of weather information• Coordination of dates and places for workshops

MONITORINGKEY ACTORS3

• Formal commitments with those organisms whose reach is all intervention area of the project in the country.

STRATEGICALLIANCES4

the products through their normal mechanisms for distribu-tion, which may include:

• E-Newsletters• Printed newsletters• Specialized magazines• Daily and weekly newspapers with national or local level reach (electronic or printed format)• Radio Programs• Television Programs• Mobile, text messaging • Website• Others

You can define a set of desirable characteristics in informa-tion products and services (WCMC 1998, CADRC 2004). In-formation products must5:

• Be aimed at specific audiences and have a purpose.• Based themselves on scientific principles and high quality data.• Be easy, as well as fast, to understand. The user-product interaction is facilitated by two features: a high level of re-presentation of objects and an intuitive interface (CADRC 2004). The user interface should be graphical in nature. Overall, the product must be easily operable, so that users can learn to use independently. However, a support system should always be available.• Be accompanied by a full survey of the sources of infor-mation and intellectual property.• Be relevant and in time for decision-making needs.• Be circulated through recognized channels.• Be available at minimal costs in time, money and admi-nistration.• Have affinities with domestic and international references.

5.4 STRATEGIC ALLIANCES WITH LOCAL AU-THORITIES

Establishing links with national/local institutions is vital for disseminating climate information. These are formal me-chanisms with operational and powerful infrastructure, es-pecially in small towns.

Local authorities are the first group that should be appro-ached and clearly and timely shown the focus of the action to be undertaken, the expected products and especially the benefits which that location will have once the action is implemented. A concise brochure and contact information is sufficient during the first approach. From then on, regular contact and good communication are important.

Having agreements with local or national entities such as municipalities or governors is very important when organi-zing advocacy and awareness activities such as workshops, press conferences or even preparing press releases. Having a cooperation agreement or letter of intent ensures com-mitment from the local authority to support such initiatives, and in turn confirms the seriousness of the partner on the quality of the type of information being broadcast to the target population of the town.

Having the support of state institutions strengthens the in-terest level of the population to participate in everything that is carried out, because it is backed by the local autho-rity.

There are also other national bodies established to channel policies for the improvement of the various development sectors. In the case of agriculture, national agricultural as-sociations, or associations of producers are key allies. Its members are farmers, community leaders or technicians, and are constantly receiving training on related issues, and undertake activities in each locality to enhance their pro-ducts and services. These networks are normally easy to reach, and therefore possess the valuable contact informa-tion and knowledge and credibility of its partners. Creating a partnership with this type of national institutions guaran-tees the mapping of players and also serves as a mean to reach a larger number of beneficiaries. The invitation to the events is done through them in places that people always gather and during the times that they know there will be a high attendance. Thus, the response from attendees is always positive and their participation greater because they are familiar with the place, the people who summon them and on a date which does not overly or interfere with their daily activities.

Thus, twenty strategic alliances were achieved with local au-thorities in the region, described in the table on next page.

5.5 THE STRATEGIC ALLIANCES WITH THE PRIVATE SECTOR

The private sector can become a great ally when creating a cooperation agreement, as it has the resources and infras-tructure needed to support various initiatives. However, it should be kept in mind in this particular category that the actions taken with them should alter their normal activities as little as possible.

Before any approach, it is important to identify the resour

5. Suárez-Mayorga A.M. (ed.). 2007. Administrator’s Guide for infor-mation on biodiversity.Biodiversity Information System of Colombia-SiB-Research Institu-te Alexander von Humboldt Biological Resources,D.C. Bogota, Colombia, 74 p.


• Ministerio del Ambiente• Municipalidad Distrital de Acolla• Municipalidad Distrital de Tunan Marca• Universidad Nacional de Huancavilca• Municipalidad Provincial de Jauja• Dirección Regional de Agricultura de Junín

PERÚ

• Asociación de Agricultores de Quilotoa• Comité de Paltas• Municipalidad de Quilotoa• Secretaría Ministerial de Agricultura de la Región de Valparaíso• Facultad de Agronomía de la Pontificia Universidad Católica de Valparaíso• Fundación de Comunicaciones, Capacitación y Cultura El Agro (FUDCA)

CHILE

• Prefectura del Departamento de La PazBOLIVIA

• Empresa Agroisleña• Asociación de Productores de Semilla Certificada de los Llanos Occidentales (APROSELLO)• Asociación de Productores del Estado de Portuguesa (ASOPORTUGUESA)

VENEZUELA

• Coorporación de Desarrollo Regional de El Oro (CODELORO)• Corporación Nacional de Agricultores y Sectores Afines (CONASA)• Municipalidad de Babahoyo• Consejo de Desarrollo del Pueblo Montubio de la Costa (CODEPMOC)

ECUADOR

ces of that company, its communication strategy and espe-cially if it performs social actions. Thus, the first communi-cation will have the following elements:

• Clear concept of what is to be achieved• Benefits to be given to the target audience• Possible mechanism to execute the action. Letter of in-tent.• Benefits for private enterprise. Recognition in terms of corporate image and social responsibility by supporting the work.

Two examples are given of successful strategic alliances with the private sector.

5.5.1 Journals Specializing in Agriculture

Upon completion of the mapping of stakeholders, we iden-tified an Industrial Group6 and an Edi-torial Group7 as potential allies, becau-se they fulfilled certain characteristics:

• Large Private Companies, easily re-cognized by the agricultural sector, their management and support.• Publish magazines focused on the topic.• Independently perform training cam-paigning on agricultural issues every year.• They have a high acceptance from population.

The first approach was to coordinate a meeting by appointment with each institution. It showed the scope of the project, the scheme

of information systems and their leading role so that the climate information reaches the end user steadily through its magazine.

In this case, during the project implementation, the NMS committed to:

1. Provide periodic climate risk maps, reports and forecasts described in a easy-to-understand language.2. Provide technical assistance to the working group for dis-semination of company information with the farmers.3. To issue at least one training workshop with company personnel on the interpretation of the technical information generated.4. Grant credit relating to the company’s products dissemi-nated through this cooperation.

The next step was to coordinate with the Department of Public Relations and Editorial on format and length of articles to be published. In this case it was easier for the company to add an article containing the following basic requirements:• Minimum length: 1 page.• Maximum length: 1 sheet.• Color maps.• Up to 3 full-color maps per page.

Subsequently early drafts of the article to be sent, in coor-dination with the MTF, were done. These products are un-derstood as information resources designed for a specific audience and defined purpose. They are the result of the compilation and presentation of analyzed or interpreted information (Villegas, Franco 2003).

Once the newsletter was ready, it was set as a template for subsequent editions. When the project was completed, this alliance passed it on to the National Weather Service, for it to maintain this operational mechanism with their products beyond the project life. There were also meetings with both institutions to establish a closer link and during the first four months, a close monitoring of the facility was done.

The newsletter is sent via email, additionally attaching as a separate files each logo and image contained in the article in the best possible resolution.

The following table describes very generally certain charac-teristics of both journals:

SPECIALIZEDMAGAZINE

DISTRIBUTIONMEDIA

METHOD OFACQUISITION

CIRCULATIONSCOPE PERIODICITY

AgroindustrialGroup

AGRIPAC

Agripac Centersof distributionof agricultural

supplies.(128 total)

FreeMagazine

5.000National Bimonthly

EditorialUMINASA

Supermarkets 3,00 USD 10.000National Monthly

6. www.agripac.com.ec7. www.elagro.com.ec

C H A P T E R V



5.5.2 Cellular Company

In the case of Ecuador, a cooperation agreement with OTE-CEL-Telefonica Movistar8 was reached to issue warning text messages or weather alerts for the entire Ecuadorian coast to the actors mapped9 by the project.

This innovative initiative is a major achievement and a mi-lestone in the use of communication technologies to distri-bute timely information to end users. Today more people use cell phones, regardless of economic status or region where they are. The mobile phone is a mass medium that reaches the user directly. Providing this service via text messages free of charge to farmers, decision makers and technicians is an effective way to communicate climate in-formation.

To achieve this, it took several meetings directly with the di-vision of Institutional Relations and Corporate Responsibi-lity of the company that was open to discuss this reporting mechanism, OTECEL-Movistar Ecuador.

The main problem they faced was related to costs. In Ecua-dor, all cellular companies work with a single company, ca-lled Message Plus10. it is responsible for transmitting text messages, it is a relay company), independent of all par-ticipants, for sending text messages. This is a very strong argument for not having the direct power to grant a free service of written messages, as the expenditure can not be assumed solely by the cell phone company, but by the relay company which is responsible for sending the text messa-ges.

To overcome this obstacle, meetings were coordinated with both the cellular company and the relay/repeater firm, leading as a concrete proposal and limited in scope with the following characteristics:

• Action Area and scope: limited. Coastal Region, 5 pro-vinces (Esmeraldas, Manabi, Los Rios, Guayas and El Oro).• Number of Members: limited. Up to 1,000 users in the provinces agreed upon.

Figure 89 shows the established mechanism of transmis-sion. At the time if INAMHI predicts a likely occurrence of any adverse event for the agricultural sector or the wider community, for example heavy rains, it will send this messa-ge via email to the repeater company, which must confirm its receipt. Then INMAHI should call the designated person to reconfirm receiving the email and find out the sending status. MPlus immediately sends that text message to the approved mobile database.

• Recurrence of transmission: limited. Only if there is proba-bility of adverse event in the action area (which is defined by the NMS).• Mechanism of Sustainability: Letter of intent signed bet-ween all stakeholders to ensure the commitment from all parties.• Message Type: According to the parameters set by the te-lephone and relay company. Warning messages of a major climate threat 11.

Once this point was reached, we proceeded to establish the mechanism for sending text messages and its format. As the repeater company is the one sending text messa-ges, it is this company that should receive the message and the database attached to the cell to send later. To reach a format, several preliminary tests between the agencies in-volved were carried out to reach a consensus.

An example of SMS message is shown below:

CELLULAR TELEPHONECOMPANY

TELEPHONESERVICE

RELAYCOMPANY

TEXT MESSAGESSERVICE

USER

8. www.movistar.com.ec 9. Users MOVISTAR mapped in 2008 in the provinces of Esmeraldas, Manabi, Los Rios, Guayas and El Oro These recipients represent far-mers’ associations, private sector, agencies and / or bodies Rescue lo-cal and sectoral authorities, community leaders and officials in NGO’S,

Figure 88. General scheme of the mechanism of telephony and text messages

IO and public institutions whose activities are aimed at risk manage-ment and disaster prevention.10. www.mplus.ec 11. No Alarm messages will be issued as only Authority local, regional or national level can create alarm messages.

C H A P T E R V


5.6 THE STRATEGIC ALLIANCES WITH THE MEDIA

We worked intensively with the media in the region, stres-sing much on the issues of sustainability and taking care in how to develop different products for each medium in ways that are as understandable and attractive as possible. A working group from each country coordinated the proper management at this crucial stage of the project, and also the research and the gathering of information to transform these newsletters or radio messages into “communicable” climate information. Also included was the background information compiled by a panel of experts in agriculture from each country, that through surveys and interviews ob-tained valuable information about traditional jargon and other cultural elements to establish communication with the target audience. We first analyzed and systematized information previously obtained on the traditional knowledge of the villagers so that they can understand the collective imagination in every sector of intervention. The results are processed in the An-nex IV. After this, and having identified the actors in the me-dia, appointments were coordinated to create partnerships for dissemination of climate information.

In this stage, there was an interaction within this group of players to receive their feedback regarding the develop-ment of formats. Because of the experience with the pri-vate sector, there was a basis for articles in newspapers and electronic magazines, but with radio it was necessary to work on the product type, then on its frequency and delivery method. At this point it is worth emphasizing the importance of having identified radio broadcast networks, and the progress that some have acquired to use the Inter-net as alternative media.

In the case of Chile, for example, we managed to imple-ment an audio narration through the radio that broadcasts Foundation of Training, Communication and Culture, an or-ganization with which, through a cooperative agreement, broadcasts over the radio the interpretation of agro-clima-tic risk map for Chile’s Region V. This big step and innova-tive mechanism for disseminating information and, in this case, the interpretation of maps, also continues to remain operational even after the project’s completion. The audios can be downloaded from FUCOA’s website: http://www.fucoa.gob.cl/radio/radio.php

1) WRITE TEXTMESSAGE

2) SENDMESSAGE

VIA E-MAIL

3) RECEPTION/CONFIRMATION

OF WARNINGMESSAGE

4) SEND TEXTMESSAGE TO

MOVISTAR USERS

Figure 89. Messaging me-chanism

Photo 1. Radio Pachaqamasa, El Alto, Bolivia.

There was also a very good approximation in Bolivia with Pachaqamasa rural radio, which has interpreters and trans-lates newsletters from the Meteorological Service of Boli-via, SENAMHI to Aymara, the language widely used in the native rural population of Bolivia.

C H A P T E R V



Figure 90. Comparative table between spoken and writ-ten language

Figure 91. List of agreements signed with media

It is no the same to write a newsletter for the radio, than to write a newsletter for the press, for the following reasons12:

That is why, when broadcasting a message over the radio, it should give the main idea in the first sentence, however, an article for print media can depict the ideas and even use more complex vocabulary.

At the regional level, agreements with several media were reached, including local newspapers, radio networks, onli-ne newspapers, weeklies and community foundations, lis-ted in figure 91.

Thus, in general, it is concluded that 06 partnerships were established with the productive sector, 16 partnerships with government institutions and 11 with media in the Andean region (Bolivia, Chile, Colombia, Ecuador, Peru and Vene-zuela) to disseminate, through different channels, climate products that were generated through the project and also the products made by each NMS.

12. Graphic CIIFEN Adapted by the book “Training Manuals” Li-brary University House Great, School of Communication, 2007.

13. Gustavo Wilchez-Chaux, 2006.14. Framework for strengthening the capacity of National Socie-ties, Colombian Red Cross.

• Radio Programas Perú RPP• Semanario Enfocando la Semana• Sin Pelos en la Pluma• Tierra Fecunda

PERÚ

• Radio emisora Nexo AM y Libra FM• El Mercurio de Valparaíso• Municipalidad de Quilotoa• Empresa Periodística El Observador• Fundación de Comunicaciones, Capacitación y Cultura El Agro (FUDCA)

CHILE

• El Diario S.A.BOLIVIA

• Editorial UMINASA• Coordinadora de Radio Popular y Educativa del Ecuador (CORAPE)• Movistar, Message Plus• Radio Naval - INOCAR y 32 Radiodifusoras• Diario La Hora de Quevedo

ECUADOR

5.7 TRAINING STRATEGIES

As a complement, we developed educational materials for community leaders, advisors of authorities and rescue groups. The compilation and systematization of similar ma-terials were the basis for developing all the parts. Without quality information there cannot be an effective participa-tion13 and this is the reason for all the effort to improve cli-mate products and services generated by the NMSs and CIIFEN, and to circulate it through the media. This was pre-sented in workshops so that the people also knows where to find weather information and how to interpret it.

In the case of Ecuador, the central concept of this material is to train potential trainers on the topic of prevention of risks to replicate basic concepts very clearly to the people in remote locations.

The material developed is an educational basic guide for risk prevention, with emphasis over the information systems implemented during the project. The material was desig-ned in such way that training activities are part of a com-prehensive training program outlined to help an individual or group to learn14. For the development of all the phases of the material, we worked in close coordination with the NMS, partner agencies and project work team to define the general outline of the educational kit. For this, we first esta-blished the general content to know what kind of activities could be done according to each chapter and the way in which this instruction would be presented. This also con-tained the experiences gained during the field trip in the design phase of the mapping of stakeholders and building strategic alliances, since they had that background on local needs and knowledge gaps.

After the completion of this stage, we obtained the general content of the guide divided into five modules:

Module I: IntroductionUse of Community Guide and training materials, how to use it.

Module II: Climate and Climate VariabilityWork table, remembering the past

Module III: Risk Management and Agro-climatic Risk Mapping.Work Table, Development of Community Risk Map

Module IV: Information and Prevention

Module V: Early Warning System To make the material user-friendly, the introduction was a module on how to handle the designed material, and even the logistical coordination that must be carried out to orga-nize a workshop.

Each chapter of the Guide has summarized and clear expla-nations through practical examples. Each chapter has su-pporting visual material to explain the central idea of that module. There are also primers for each chapter, promo

SpokenLanguage

WrittenLanguage

It is moredisarranged

It is morepersonal

It is moreexpressive

It is moreimpersonal

It is moreaccurate

It is moresorted

C H A P T E R V


ting group activities during the workshop. A cartoon was also created to conclude each workshop, which had a simi-lar version to a printed booklet that was given along with the other printed materials. All this material is supplemen-ted to provide the training workshop, and to have visual su-pport, it is not imperative to have an In Focus projector for the workshop; it can even be arranged without a computer. This effort integrated the support from other cooperating institutions, such as the ProVention Consortium, and DIPE-CHO, via the V action plan.

The items that make up this training kit are:

• Community Preparation Guide: Guide notebook for the trainer; it contains the methodology to organize works-hops, the members who must attend, place, time, develo-pment of each module, management and guidance during the work tables.• Activities Primers: Direct complementary elements; the guide, they give instructions on the activity to be develo-ped, and also how to develop the activity.

• ”El Temporal” newspaper: Newspaper that concisely summarizes the core concepts of each module in the guide. Element of consultation at any time during the workshop.

• Cartoon: Animated story about disaster prevention. Additional element to reinforce concepts

• Pamphlet: Brochure with additional tips on how to take care of our environment.

• Lunar Calendar: Given at the workshops, wall calendar.

• Pocket guide: List of provincial emergency telephones.

Besides the items mentioned above, participants were gi-ven folders and pens and material required to develop the activities.

Overall, the training workshops had an introductory phase before the development of the topics. The development and implementation of information systems was explained.

Then an introduction was given on basic concepts to fami-liarize the participants on the methodology and use of wor-kshop tools. During the workshops, participants applied the concepts through group activities, sharing experiences to synthesize them on a chart about a locally-impacting adverse climatic encounter (work table module II) and to develop a climate risk map of their sector, which identified risk zones, vulnerable areas, risk areas, and proposed pos-sible temporary shelters and evacuation routes (work table module II).

At the end of each workshop, each attendee was given a certificate, a community guide and printed visual support material, including primers for the work tables contained in a folder for 10 people. That is, each attendee received a community guide and 10 folders so that they can replicate the workshop in their locations. It is important to point out that it is not necessary to have projector or computer to implement the workshop.

C H A P T E R V



In the case of the other Andean countries, we worked in close coordination with each NMS to develop training materials, answering the needs of each intervention area, and according to the capabilities identified during the ma-pping. Special emphasis was given to the interpretation of agro-climatic risk maps and to each NMS newsletters, which were improved by the project team.

The workshops ensured the presence of the media, local authorities and were attended by representatives of trade associations, representatives from disaster management agencies, rescue agencies, production houses, technicians, consultants from local authorities, community leaders, among others.

National Workshop on PROSUKO facilities. Community Pucarami, Bolivia.

National Workshop. The Ligua, Valparaiso Region, Chile

National Workshop. Aragua, Venezuela

National Workshop. Huancayo, Junin Department, Peru

Brunildo and Magola, characters of the serie “Let`s un-derstand wheater to can live with him”

C H A P T E R V


CHAPTER VICapacity building in

Western South America

6.1 REGIONAL WORKSHOP TRAINING ON CLIMATE MODELING STATISTICS

The Regional Training Workshop on Climate Modeling Sta-tistics, took place on October 8-13, 2007 at the Bolivarian Military Aviation’s Weather Service installations in Maracay, Venezuela. The workshop gathered 18 people from the National Meteorological Services of Bolivia, Chile, Colom-bia, Ecuador, Peru and Venezuela. The workshop included the participation of two regional trainers: Ángel Muñoz (Venezuela) and Marco Paredes (Peru), who combined the theoretical and the practical phases of the course that was based on the implementation of the CPT tool (Climate Pre-dictability Tool) developed by the IRI. The workshop had two components: instructional and applicability. The results of the workshop were: The seasonal forecast for the Octo-ber to December 2007 quarter, with actual data for each country prepared by each participant, the monthly and bi-monthly forecast for each country presented individually, the Quarterly regional seasonal forecast and preparation of a document discussed by the participants on methodo-logical principles and recommendations for application and implementation of Climate Modeling Statistics in the region.

Photo of the participants of the Regional Workshop on Numerical Modeling Statistics (Maracay, Venezuela, Octo-ber 8-13, 2007)

Participants of the Regional Workshop on Numerical Mode-ling I (Lima, 19 to 24 November 2007)

Training by technicians participating NMHSs in the use of CMM5 and CWRF

6.2 REGIONAL TRAINING WORKSHOP ON NUMERICAL MODELING FOR CLIMATE PRE-DICTION

The Regional Training Workshop on “Modeling Climate Statistics” was held on November 19-24, 2007 at the head-quarters of the Meteorology and Hydrology Service of Peru (SENAMHI) in Lima. The workshop gathered 14 people from the NMHSs of Bolivia, Chile, Colombia, Ecuador and Peru. The lecture, exercises, methods and practice sessions were conducted by D. Angel G. Munoz Solorzano, profes-sor at the University of Zulia (Venezuela) and Deputy Direc-tor of Scientific Modeling Center (CMC).

6.3 REGIONAL TRAINING WORKSHOP FOR AGRO-CLIMATIC RISK MAPPING

The International “Methodology for Agro-climatic Risk Ma-pping” Workshop took place on January 14-19, 2008 at the Santiago de Guayaquil Catholic University in Guayaquil, Ecuador. The workshop included six people from the Na-tional Meteorological Services of Bolivia, Chile, Colombia, Ecuador, Peru and Venezuela, eleven people representing government agencies and private institutions of Ecuador (INAMHI, INOCAR, SENPLADES, UCSG, MAGAP, MAA, Cedega). The workshop had the participation of: Angel Llerena, Harold Troya and Nadia Manobanda, who combi-ned the theoretical and practical phases of the course that was based on the explanation of the methodology and risk mapping for the agricultural sector. The workshop had two components: instructional and applicability. In addition, with the participation of Juan Jose Nieto in the operation of the Surfer tool for weather forecast mapping. The results of the workshop were: development of an agro-climatic risk map, with real data prepared by groups of four participants and a group presentation by participants on the methodo-

CHAPTER VI



C H A P T E R VI

Photo of the participants of the Regional Workshop on Numerical Modeling of Weather and Climate II May 26-31, 2008

logical principles and applications learned along with re-commendations for extension, dissemination and imple-mentation of the methodology to local and regional level.

6.4 REGIONAL WORKSHOP ON NUMERICAL MODELING OF WEATHER AND CLIMATE II

The Regional “Numerical Modeling of Weather and Cli-mate” Workshop was held on May 26 -31, 2008 at Escuela Politécnica del Litoral (ESPOL) in Guayaquil, Ecuador. Wor-kshop participants included five people from the National Meteorological Services of Bolivia, Chile, Ecuador, Peru and Venezuela as well as eighteen people representing go-vernment agencies in Ecuador (INAMHI, INOCAR, ESPOL, Institute of Fisheries). The workshop included the participa-tion of Prof. Angel G. Muñoz S. (Center for Scientific Mode-ling, CMC, University of Zulia - Venezuela) as an instructor, who introduced the content of theory and practice ses-sions: the former focused on atmospheric-oceanographic phenomena and their involvement in the weather and cli-mate forecast, as well as the related physical mathematical fundaments, global and regional models, downscaling and validation. In the practice sessions, attendees were able to compare in detail the differences between models and ob-servations, and carry out their own executions in (C) MM5 and (C) WRF weather and climate models, fed with data from the NNRP, GFS and CAM model, from which CMC is in operational mode to make regional forecasts. Finally, it is worth noting that the workshop allowed to formally esta-blish the Regional Modeling Group (MRG), which includes all the participants of the meeting and had the support of the National Meteorological Services.

6.5 INTERNATIONAL TRAINING WORKSHOP ON CLIMATE DATA PROCESSING

The workshop was organized by CIIFEN and the Venezue-lan Aviation Weather Service (SEMETFAV) on October 6-7, 2008, in which the participants learned techniques for pro-cessing, filtering, quality control and time series standardi-zation. They reviewed theoretical concepts as well as some tools (codes in matlab) for such purposes.


CHAPTER VIIPerformance indicators

BOLIVIA

In the Department of La Paz-Northern Altiplano, Oruro-Central Altiplano, Potosi, Southern Highlands, Bolivia, surveys were conducted on the Agrarian Union, local mu-nicipal governments, small farmers, local NGOs, lender ins-titutions, Ministry of Agriculture (SIBTA), Ministry of Lands, Ministry of Planning, as a baseline which gave the result that only 2% has access to climate information, while the remai-ning 98% does not have access to it. This information was selected as a baseline by the project team and SENAMHI to disseminate weather products mainly through newspa-pers and radio. Once the entire project management and national workshops were completed in the Department of La Paz, a new survey was conducted focusing on the people who did not receive information. The results showed that of the 98% that at the beginning of the project did not have access to climate information, 5% now had access to climate information. In addition, out of the total number of people surveyed, 74% use climate products for the agricul-tural management of their crops to varying degrees.

CHILE

In Region V, Chile, initial surveys were carried out and con-cluded that 62% has access to climate information, while the remaining 38% does not have access to it. This infor-mation was taken as a baseline for the project team and DMC to disseminate weather products mainly through newspapers, radio and internet. Once the work and natio-nal workshops were completed in Region V, a new survey was conducted focusing on the people who did not recei-ve information. The results showed that of the 38% that at the beginning of the project said they did not have access to climate information, 4% now had it. In addition, out of the total number of respondents, 67% find a high applica-bility of climate products for the agricultural management of their crops. A question regarding the major sources of climate information was also asked; the newspaper is the largest source with 25%, followed by email with 21% cell phones with 20%, 18% radio and internet with 16%.

COLOMBIA

In the Department of Tolima and Sabana de Bogotá, Co-lombia, a baseline survey was undertaken on 26 flower-producing and rice-producing companies in the Savannah Bogota and in central Tolima, respectively. This number constitutes 10% of all companies (260) in the region. The results showed that 60% has access to climate information while 40% does not have access to it. This information was taken as a baseline by the project team and IDEAM. At the end of the project and upon completion of the natio-nal workshops, a new survey was conducted focusing on the people who did not receive information. The results showed that of the 40% that at the beginning of the pro-ject said they did not have access to climate information, 33% now had it. In addition, out of the total number of respondents, 70% found climate products very useful and applicable for farm their crop management. A question re-garding the main sources of climate information was also

asked, IDEAM being the largest source with 47%, followed by FEDEARROZ with 38%, 10% and Internet radio 5%.

ECUADOR

In the provinces of Manabi, Los Rios and Guayas, Ecuador, initial surveys were conducted which concluded that only 4% has access to climate information, while 96% does not have access to it. This information was taken as a baseline for the project team and INAMHI to disseminate weather products through newspapers, radio, internet, email and private sector journals. Once the project management team and national workshops in the 3 provinces were finis-hed, a new survey was conducted focusing on the people who did not receive information. The results showed that of the 96% that at the beginning of the project said they did not have access to climate information, 94% now had access to it. In addition, out of the total number of people surveyed, 87% finds a high applicability of climate products for the agricultural management of their crops. A ques-tion regarding the major sources of climate information was also asked and the newspaper is largest source with 41%, followed by radio at 32%, magazines with 18% and 9% cell phones.

PERU

In the cities of Cockaigne and Huancayo in the Mantaro Va-lley, Junín, Peru, initial surveys were conducted which con-cluded that only 6% has access to climate information while the remaining 94% does not. This information was taken as a baseline by the project team and SENAMHI to dissemi-nate weather products through newspapers, radio, internet and email. Once the project team and national workshops in the two cities mentioned above were finished, a new survey was conducted focusing on the people who did not receive information. The results showed that of the 94% that at the beginning of the project said that they did not have access to climate information, 15.98% had access to it now. In addition, of the total number of those surveyed, 83% finds a high applicability of climate products in the agricultural management of their crops.

VENEZUELA

In the towns of Turén, Acarigua and Guanare in Portugue-sa State, Venezuela, the surveys conducted as a baseline for the project team concluded that 43.3% has access to climate information, while 56.8% does not. This informa-tion was taken as a baseline for the project team and SE-METAVIA in order to disseminate through internet, email and associations climate products to end users. Once the project management team and national workshops were finished in these three cities, a new survey was conducted focusing on people who did not receive information. The results showed that of the 56.8% that at the beginning of the project said that they did not have access to climate information; 45.55% now had access to it. Furthermore, of the total number of those surveyed, 85.67% finds a high applicability of climate products into the agricultural mana-gement of their crops.

CHAPTER VII



Power, Scott, Plummer, N., Alford, P. 2007. Making Climate Model Forecast more useful. Australian Journal of agricultural Research. 58: 945-951.

Power, Scott, Sadler, Brian, and Nicholls, Neville. 2005. The Influence of Climate Science on Water Management in Wes-tern Australia Lessons for Climate Scientists. BAMS-86-6-839.

Santillán, G., 2005. Manual para la Prevención de Desastres y Respuesta a Emergencias. La experiencia de Apurímac y Aya-cucho. ITDG.

Randall E et al. A beginner’s guide to structural equation modeling pg. 38.

Suárez-Mayorga A.M. (ed.). 2007. Guía del administrador de información sobre biodiversidad. Sistema de Información sobre Biodiversidad de Colombia -SiB-, Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Bogotá D.C., Colombia, 74 pp.

Sutherland, W., et al. 2009. One Hundred Questions of Importance to the Conservation of Global Biology Diversity. Society for Conservation Biology. 1523-1739.

Tarbell, T.C., T.T. Warner and R.A. Anthes, 1981. The initialization of the divergent component of the horizontal wind in mesoscale numerical weather prediction models and its effect on initial precipitation rates. Mon. Wea. Rev., 109, 77-95. UNIDR, 2009. Terminología sobre Riesgo de Desastres.

United Nations Organization, 2001. Tools to Support Participatory Urban Decision Making Process: Stakeholder Analysis. Urban Governance Toolkit of HABITAT program.

United Nations, UNEP, 2007. Biodiversity and Climate Change. International day for Biological Diversity. 1-48.

Villagrán De León, J., 2006. Vulnerability. A conceptual and Methodological Review. United Nations University Institute for Environment and Human Security. Series UNU – EHS No. 4.

UNISDR, 2009. Terminología sobre Reducción de Riesgo de Desastres 2009 para los conceptos de Amenaza, vulnerabilidad y riesgo.

Wilches-Chaux, G., 2007. Conceptos Básicos sobre Gestión de Riesgo y Seguridad Territorial.

Wilches-Chaux, G. Brújula, Bastón y Lámpara para trasegar los caminos de la Educación Ambienal. Ministerio del Ambiente, Vivienda y Desarrollo Territorial, República de Colombia.

Xiao, Q., W. Guo, and X. Zhou, 1996. Preliminary results from numerical experiments of a heavy rain process with PENN STATE/NCAR MM5, Advance in Atmospheric Sciences, 13(4), 539-547.

William H. Press. Numerical recipes: the art of scientific computing pg. 349.

Yucel, I., W. J. Shuttleworth, X. Gao, and S. Sorooshian, 2003. Short-Term Performance of MM5 with Cloud-Cover Assimi-lation from Satellite Observations, Monthly Weather Review, 131, 1797-1810.

BIBLIOGRAPHIC REFERENCES



C H A P T E R VII

REGIONAL INDICATORS PROJECT INTER-VENTION AREAS

Baseline surveys developed during the first phase of the project indicate that the percentage of population with information were: Bolivia 2%, Chile 62%, Colombia 60%, Ecuador 4%, Peru 6% and Venezuela 43.2%. At the con-clusion of the project, surveys indicate that people with information were: Bolivia 6.9%, 63.52% in Chile, Colom-bia 73.2%, Ecuador 96.24%, Peru 21.03% and Venezuela 69.08%.

This represents an increase of new users in Bolivia by 4.9%, 1.52% in Chile, 13.2% in Colombia, 92.94% in Ecuador, in Peru 15.03% and 25.88% in Venezuela. The increase of users in the region is 25.58% and the population that finds climate information applicable is 77.78% in the Andean region.

% A

vera

ge

% A

vera

ge

Vene

zuel

a

Vene

zuel

a

Peru

Peru

Ecua

dor

Ecua

dor

Colom

bia

Colom

bia

Chile

Chile

Bolivi

aBol

ivia

% Population with information, beginning of project

% Increase in New Users Population that finds infor-mation applicable

% Population with information, end of project


CHAPTER VIIILearned Lessons

IMPLEMENTATION OF THE VIRTUAL CORE OF CLIMATE APPLICATIONS (VCCA)

• The project component of the Core Development Virtual Climate Applications in the developmental stage passed through several obstacles that were part of the software de-velopment cycle; however, it was possible to identify three lessons:

• The development of VCCA was a team effort, and although It was coordinated over distance, it had positive results. The join of Andean countries on this effort and the creation of an integrated database showed that it is possi-ble to create regional products and make them available to the general public, despite distances and differences.

• Creating high quality software applications and availabi-lity are feasible on Open Source tools, without detracting functionality in the least. These are a highly reliable alterna-tives to ensure the permanence of the application in time.

• The rapprochement between the Web-based computer tools and the end user is usually possible, offering easy to use products that provide useful information for the activi-ties of these users.

IMPLEMENTATION OF STATISTICAL MODELS FOR CLIMATE PREDICTION

• Among the problems identified it can be mentioned that upon not having a standardized methodology, the different national contributions to regional prediction were not uni-form and sometimes physically inconsistent, having results diametrically opposite in neighboring countries along their borders. Moreover, the statistical background on the me-thodology of the terciles, correlations and linear combina-tions seems to be unfamiliar to some of the users of the program. For this reason their climatic perspectives were sometimes based on subjective evaluations. The cause of this seems to be that the resources of some institutions are limited to daily tasks and they can allocate very little of their budgets to research and training. So far there have been few but valuable training opportunities on these concepts to the participants in these forums. CIIFEN has made great efforts to disseminate a set of executable programs that utilize user-friendly graphic interfaces to be used by meteo-rological services in weather forecast.

• One of the biggest challenges was the standardization of knowledge related to the use of software available in the world; this activity was a gradual process in the field of sta-tistical modeling that was done by itinerant experts and can be synthesized in 3 steps:

• 1st Stage: Use of Exever (program developed by NOAA/ OGP), and promoted by CIIFEN in western South Ameri-ca in 2004 as a climate prediction tool, which sought linear correlations between two variables for a particular season, the predictor and predicting, through cross-correlations -1 and 0 lags (monthly time scale). This software was easy to use in areas where there weren’t many stations; however, it

entailed more time when many stations worked because the runs were carried out station by station and variable by variable.

• 2nd Stage: of CPT (program developed by IRI), and pro-moted by CIIFEN in western South America in 2005 as a climate prediction tool, through which was determined the maximization of the linear correlations between a set of predictor variables and a set of one variable that was loca-ted over a predicting area. This stage had a breakthrough with respect to numbers of processing stations. The use of predictors and access of data to be used as predictors via the IRI data library was an important starting point for the use of CPT.

• 3rd Stage: Use of the CPT as a validation tool in weather forecasting; this stage sought to exploit the validation com-ponent that CPT has immersed after obtaining preliminary results.

• Each stage was reinforced with permanent itinerant visits to each of the participating countries and training conduc-ted within the Climatic Forums of western South America, which is approved by the World Meteorological Organiza-tion coordinated by CIIFEN.

• The use of CPT achieved the standardization of me-thodologies for climate prediction, obtaining coherent and comprehensive (regional) results of precipitation and tem-perature variables. The delivery time of forecasts to the population was summarized and the search for shorter tem-porary horizons (semester and monthly scales) was begun.

• The seasonal forecast for western South America, based on CPT as a common tool, is an operational product that is generated monthly and distributed to thousands of users in the region.

IMPLEMENTATION OF NUMERICAL MODELS FOR CLIMATE PREDICTION

• The attention and motivation on the part of NMHSs to this activity was always very high.

• An interesting aspect was that although the original ob-jectives envisioned only the start of experiments in downs-caling in retrospect and with only a single regional model, it was possible, working in coordination, to install and con-figure in the vast majority of countries two regional models (CMM5 and CWRF) and additionally, integrations of the models in experimental forecasting was begun.

• In some NMHSs, the Internet connection was slow, so the decision to back up everything necessary on a portable hard drive was correct.

• Although the experiments were completed, a pending as-signment of each NMHS involved obtaining each model’s climatology. Indeed it is a task that, while not complicated, involves a long period of computing time to reach its end and that is related to the availability and computational ca-

CHAPTER VIII



C H A P T E R VIII

pabilities present in each Weather Service and associated with the Project. At present some countries are completing this assignment, but others have not continued it.

• Two training workshops were conducted: One (November 2007) regarding installation, configuration and basic execu-tion of the CMM5 model. The second (June 2008) focused on deeper aspects of the climatic configuration and imple-mentation of both CMM5 and CWRF. It is considered that a single workshop definitely was not enough, so that having organized a second to reinforce ideas and methodologies was worthwhile. Conducting two training sessions a year of such topics is a recommended idea.

• Having articulated training workshops, with specific tasks in each NMHS and the corresponding technical support contributed to the development of capacities for numerical modeling of the NMHS in the six countries.

IMPLEMENTATION OF AGRO-CLIMATIC RISK MAPS

• The development of agriculture climate risk maps invol-ves interdisciplinary development that brings together cli-matologists, geographers, agronomists, sociologists and computer programmers in a common discussion, it cannot be done unilaterally.

• One of the key steps was the development of the agricul-tural risk conceptual model. Despite the stringency of the definitions it has to take into consideration the feasibility of obtaining the information, the scale and the accuracy of the information. In this sense, the quality of the final infor-mation is not a function of the complexity and number of variables involved but on the strength and availability of the variables to be used.

• The development and validation testing involved was ba-sed on experience from experts and users.

• For this type of implementation, it would be necessary two workshops, one for discussion and conceptual model validation and one for training in the design of the GIS.

• Support from the project team through itinerant missions in each country, was critical to its implementation.

• Such tools require a specialized unit NMHSs counterpart, especially for its sustainability and continuous improve-ment.

IMPLEMENTATION OF LOCAL SYSTEMS OF CLIMATE INFORMATION DISSEMINATION.

• The technical language of climate information is difficult to convert. A native language like Aymara requires further work. It should be linked with cultural elements of ancient knowledge of the climate.

• Working steadily with services throughout the process helped to see the need for additional efforts to establish protocols for dissemination of products and services in con-sensus with everyone and not wait for someone to appoint a service technician to design it.

• The local team in each country should have worked lon-ger on the project to build alliances, at least 12 months of hard work are needed.

• The development of baseline and final project measuring surveys should have been carried out by a single group of consultants for a more detailed analysis of the impact of the results and focusing on the social, cultural and political reality of each country in such a manner that it is easy to compare it with other countries in the region.

• The work to create alliances to form the distribution net-work required more time, money and manpower in each country. In fact this only warranted a separate project due to its complexity and especially because the articulation of people, institutions and other organizations demands time, building of trust, face to face contact and lots of patience.

• The ideal time to give workshops on prevention and pre-paredness in hydro-meteorological issues is before the ra-iny season.

• During the development of training material, the obser-vations made by of representatives of each group about the users maps are important to strengthen the structure of content.

• After having a mapping of actors, it would be appropria-te to identify some important characteristics of the actors, such as having influence over another group of actors, affi-nity with the topics covered, level of cooperation or if they are active in the field where they work, to name a few. With this you could construct a flowchart that represents the real inter-agency relationships.

• It takes time to accept changes in practices and behavior of users due to the use of new technologies.

• The reliability of some development sectors in climate fo-recasting is still very limited.

• Few consecutive incorrect forecasts cause the people to distrust research centers.


CHAPTER IXFuture Actions


Actions to be developed according to each application are outlined:

Regional Climate Database

• Update information by adding new records, according to the Data Access Protocol signed by the National Meteoro-logical Services.

• Application of statistical analysis techniques for data qua-lity.

• Addition of monthly data on precipitation, maximum and minimum temperature.

Map Server

• Update information by country, including forecast climate information and crop types.

• Encourage the use of Open Source tools for the genera-tion of SIG products.

Climate Modeling Products Viewer

• Publication of forecast products for different areas and with different climate models.

Virtual Library

• Addition of new publications for general consultation, de-pending on their availability..


• It is necessary to continue to coordinate activities for stan-dardization of validation and verification criteria of seasonal forecasts. This is a process that will take time and the results obtained will serve to guide the efforts of climate forecas-ting in the short and medium term.

• Have virtual conferences on a monthly basis among the participants, guided by regional experts or CIIFEN.

• Share and expand this process to other regions of Latin America and the Caribbean.


It is the opinion of the National Meteorological Services (NMSs) that the organizational and structural initiative that currently exists for implementing regional climate models, configured with the settings chosen on the basis of the ex-

perience and knowledge of the local experts of each coun-try, can continue. In this sense, CMC and CIIFEN have conti-nued joining forces, along with NMSs and universities in the Andean region (Universidad del Zulia, Universidad Católica de Bogotá, Universidad Mayor de San Andrés, Universidad de Chile), constituting what has been called the “Extraordi-nary Events Andean Observatory.

The Center’s purpose is to articulate institutions, techni-cians and technological resources to provide scientific tools which, through climate forecast, assist in decision making, the creation of early warning systems and risk management within the framework of the Andean region.


• To continue improving the methodology for the agricul-ture risk estimation.

• Incorporate information from remote sensing related to water retention capacity of soil, vegetation index and other variables.

• Migrate the current system entirely to Open Source.

• Incorporate the new GIS spatial analysis tools.

• Disseminate the methodology to other countries.

IMPLEMENTATION OF LOCAL SYSTEMS OF CLIMATE INFORMATION DISSEMINATION

• Replicate the experience with the companies of cell pho-nes and repeater in Ecuador and other countries.

• Expande the space in the journals with which there are cooperation agreements, to provide a space with informa-tion targeted specifically to the rural community.

• Increase the number of partners in private companies to disseminate information.

• Develop and disseminate spots containing warnings and short messages over the radio and certain television spa-ces.

• Replicate the most relevant experiences in the region to create pilot projects in other areas.

• Maintain contact with the mapping of participants esta-blished through personal visits, email, local workshops, and videoconferences, among others.

CHAPTER IX



CHAPTER XElements of Sustainability


• The NVAC, since its initiation, was designed as a sustaina-ble system in time, at the stages of development, mainte-nance and updating. The following elements of the sustai-nability of the system have been defined:

• Open Source Architecture Software: To ensure sustaina-bility of the NVAC, from the planning stage, the use of soft-ware under Open Source license was determined. Under this philosophy, future license renewal through additional payments is avoided. Respecting this principle, each of the components of NVAC as well as the applications running on it, have been developed using Open Source tools, thus ensuring the retention and upgrading of applications on time.

• Free Access: Access to the NVAC applications is free, any user with an Internet connection can access, view and ob-tain information.

• Information Update: Authorized users of each NMS are able to update the weather information as it is generated; this way, products are guaranteed to have the latest infor-mation gathered by regional NMHSs.


• The regional project helped to consolidate the formation of an important group of techniques from the six countries of the region, with the capacity to expand the critical mass of people that successfully operate, understand and apply the CPT.• CIIFEN, as an international organization with strong ties to NMHSs, continues to promote improved seasonal fore-casting and training and technical assistance opportunities.


• During project implementation, CIIFEN signed a coope-ration agreement with the Center for Scientific Modeling of the University of Zulia.

• This partnership has ensured the important technical su-pport of CMC on numerical modeling, within the various lines of work that CIIFEN maintains with NMHSs in the re-gion.


• CIIFEN maintains a technical support to all NMHSs regar-ding risk maps to ensure sustainability.

• Manuals produced and distributed to NMHSs, allow to work in the GIS and do subsequent alterations.

IMPLEMENTATION OF LOCAL SYSTEMS OF CLIMATE INFORMATION DISSEMINATION

• One of the positive factors identified has been the level of commitment reached by all NMHSs, derived from a cohe-rent Operations Plan and focused on the institutions.

• The mechanisms implemented to disseminate informa-tion through private enterprise and the media were given to the NMHSs to adopt them as their own and integrate them into the mandatory activities of the Service. This new window for the NMS is an opportunity to position the ser-vice as a good source of products and services developed specifically for the end user.

• Having private enterprise as an ally ensures a more pro-longed duration of an agreement since it is more stable than political positions. Having made agreements with this group gives greater strength to the sustainability of what has been implemented.

CHAPTER X



Aldrian, E., 2009. Regional Climate Information for Risk Management (needs). World Climate Conference 3, white paper Nº12.

Andrade, A. (ed), 2007. Aplicación del Enfoque Ecosistémico en Lationoamérica. CEM-UICN. Bogotá, Colombia. 1-89. Anthes, R.A. and T.T. Warner, 1978. Development of hydrodynamic models suitable for air pollution and other mesometeo-rological studies. Mon. Wea. Rev., 106, 1045-1078.

Brown, C. and Hansen, J.W., 2008. Agricultural Water Management and Climate Risk. Report to the Bill and Melinda Gates Foundation. IRI Tech. Rep. No. 08-01. International Research Institute for Climate and Society, Palisades, new York, USA. 19 pp.

Cardona, O., 2005. Indicadores de Riesgo de Desastres y de Gestión de Riesgos. Banco Interamericano de Desarrollo, De-partamento de Desarrollo Sostenible, División de Medio Ambiente. 1-53.

Colle, B. A., and C. F. Mass, 2000. The 5-9 February 1996 flooding event over the Pacific Northwest: sensitivity studies and evaluation of the MM5 precipitation forecasts. Mon. Wea. Rev, 128, 593-617.

Colle, B. A., C. F. Mass, and D. Ovens, 2001. Evaluation of the timing and strength of MM5 and Eta surface trough passages over the Eastern Pacific. Weather and Forecasting, Boston, MA, 16(5): 553-572.

ESPOL, 2009. Aplicación de Técnicas SIG, Sensoramiento Remoto y análisis multitemporal para la detección de Impactos Antropogénicos en la cobertura de suelos y su proyección para el 2010. Caso de Estudio: Estuario de Santos (Brasil). Revista Tecnológica ESPOL. 1-7.

Fernández, M., 1996. Ciudades en Riesgo, Declaración Ambiental, Riesgos Urbanos y Desastres. LA RED.

Ferradas, P., 2006. Metodologías y Herramientas para la Capacitación en Gestión de Riesgo de Desastres. INDECI-Instituto Nacional de Defensa Civil, Soluciones Prácticas-ITDG.

Geraldo, M., 2003. Hacia una Cultura de Prevención de Desastres, desde la Educación. Memoria del Seminario Taller Inter-nacional sobre Prevención y Atención de Desastres en la Educación, Lima, Perú.

Kaipi, K., et al. 2005. Gestión de Riesgo de Amenazas Naturales en Proyectos de Desarrollo. Serie de Informes de Buenas Prácticas del Departamento de Desarrollo Sostenible, Banco Interamericano de Desarrollo. 1-55.

Marti Ezpeleta, A., 1993. Cálculo del Riesgo de Adversidades Climáticas para los Cultivos: Los Cereales de Verano en Mon-tenegros. Dpto. de Geografía y Ordenación del Territorio, Universidad Zaragoza.

Martínez, R., 2008. Users Liaison in RCOF Position Paper. International Expert Review Meeting on Regional Climate Outlook Forums, Arusha, United Republic of Tanzania.

Martínez, R., 2009. Regional Climate Information for Risk Management (capacilities). World Climate Conference 3.

Maskrey, A. (ed). 1998. Navegando Entre Brumas, La aplicación de los Sistemas de Información Geográfica al Análisis de Riesgo en América Latina. LA RED. 1-11.

Martínez, Rodney, 2006. Gestión de la información y servicios de predicción climática para la reducción de impactos en el sector agrícola en Sudamérica. Campiñas, 8-15.

Miao, J.-F., D. Chen, and K. Wyser, 2006. Modelling Subgrid Scale Dry Deposition Velocity of O3 over the Swedish West Coast with MM5-PX Model. Atmospheric Environment 40 (3): 415-429.

Oficina de los Estados Unidos de Asistencia para Desastres en el Exterior, USAID, 2007. Evaluación de Daños y Análisis de Necesidades Nivel Toma de Decisiones.

Michalakes, J.; Dudhia, J et al. 2004. The weather Research and forecast model: Software architecture and performance. 11th ECMWF workshop on the Use of High Performance Computing in Meteorology, Reading U.K.

Pereira Fo., Augusto J. Crawford, Kenneth C. Stensrud, David J., 1999. Mesoscale precipitation fields. Part II: Hydrome-teorologic modeling. Journal of Applied Meteorology, Boston, MA, 38(1): 102-125.

BIBLIOGRAPHIC REFERENCES


Technical Guide of a Regional Climate Information System

Documents

c i n e

r e s e

e n t i n t h e

e r i c

r c h c e n t r e o

r c h c e n t e r o

r e g i o n

r e n h