LIMPOPO RIVER BASIN - World Meteorological · PDF fileSADC–HYCOS at Department of Water Affairs, South Africa; ... Limpopo River Basin was prepared for the World Meteorological...

LIMPOPO RIVER BASIN

A PROPOSAL TO IMPROVE

THE FLOOD FORECASTING and

EARLY WARNING SYSTEM

World Meteorological Organization

2

© 2012, World Meteorological Organization

i

Table of Contents ACKNOWLEDGEMENT ...............................................................................................................iii

ACRONYMS AND ABBREVIATIONS ..............................................................................................iv

EXECUTIVE SUMMARY...............................................................................................................vi

Chapter 1 : BACKGROUND............................................................................................................. 1

1.1 LIMPOPO RIVER AND FLOODS............................................................................................... 2

1.1.1 Hydrology and Floods of Limpopo River ......................................................................... 3

1.1.2 Floods Monitoring and Information Management........................................................... 5

1.2 FLOODS AND DISASTER RISK MANAGEMENT COOPERATION ......................................................... 8

1.2.1 Food Disaster Stakeholder Analysis................................................................................ 8

1.2.2 Regional Policy, Legal and Institutional Arrangements..................................................... 9

1.2.3  Flood Disaster Relief Operations and Regional Cooperation............................................11

1.3 Flood Risks Management Challenges.............................................................................11

1.3.1 Inadequate Water Resources and Flood Monitoring Systems .........................................12

1.3.2 Limited Data Exchange and Technical Cooperation ........................................................12

1.3.3 Uncoordinated and Incomplete Forecasting and Warning Systems .................................12

1.3.4 Limited Institutional and Capacity Development............................................................13

1.3.5 Summary of Forecasting Challenges and Conclusions .....................................................14

Chapter 2 : REVIEW OF FLOOD FORECASTING AND WANRING SYSTEMS...................................15

2.1 EXISTING FLOOD FORECASTING DEVELOPMENT IN LIMPOPO RIVER BASIN .......................................15

2.1.1 Southern African Region Flash Flood Guidance System...................................................15

2.1.2. Severe Weather Forecasting System for Southern Africa. ...............................................16

2.1.3 SADC‐HYCOS Telemetry System....................................................................................17

2.1.4 South Africa River Flow Flood Forecasting Systems ........................................................21

2.1.5 ARA Sul Limpopo Flood Flow Flood Forecasting System..................................................22

2.2 MODEL FLOOD FORECASTING AND EARLY WARNING SYSTEMS.....................................................27

2.2.1 Real Time Flood and Related Data Monitoring...............................................................27 

2.2.2 Flood Forecasting Models.............................................................................................28

2.2.3 Flood Early Warning and Response ...............................................................................31

2.2.4 Trans‐boundary FFEWS Management Practices .............................................................34

Chapter 3 : RECOMMENDATIONS FOR IMPROVED LIMPOPO FFEWS ........................................37

3.1 IMPROVEMENT OF LIMCOM MONITORING SYSTEMS................................................................37

3.1.1 Consolidation and Modernizing Telemetry Network.......................................................37

3.1.2 Harmonizing Data Collection and Processing Routines ...................................................38

ii

3.1.3 Development and Launching Data Sharing Systems........................................................38

3.1.4 Trans‐boundary Telemetry Network Management.........................................................39

3.2 CONSOLIDATION OF RIVER FLOW FORECASTING AND EARLY WARNING...........................................39

3.2.1 Objectives and Outputs of the River Flow Forecasting System ........................................39

3.2.2 Establishment of Regional River Flow Forecasting Centre...............................................40

3.2.3 Development/Commissioning River Flow Forecasting Models ........................................40

3.2.4 Manual of Standards and Guidelines of Limpopo RFEWS................................................41

3.3 RFEWS INSTITUTIONAL DEVELOPMENT AND CAPACITY BUILDING.................................................42

3.3.1 Technical Capacity   Building Programme Development..................................................42

3.3.2 Capacity Building and Training......................................................................................42

3.4 WORK PROGRAMME, INPUTS AND COST ESTIMATES .................................................................42

3.4.1 Work Programme and Input Requirements ...................................................................42

3.4.2 Budget and Cost Estimate of the RFEWS........................................................................44

REFERENCES ............................................................................................................................47

List of Tables Table 1-1  Some Catastrophic Floods and Droughts in Southern Africa .......................................... 3 Table 1-2  National Hydro-meteorological Network in Limpopo River.............................................. 7 Table 1-3  Public Stakeholders of Limpopo FFEWS ...................................................................... 8 Table 2-1  Limpopo Basin SADC-HYCOS Telemetry Network Station List .....................................19 Table 2-2  Limpopo FFEWS Telemetry Network in Mozambique...................................................23 Table 2-3  Limpopo FFEWS Cross-section Location and Distances Apart .....................................26 Table 2-4  Table of MRC Flood Risk Management Strategies and Priorities...................................36 Table 3-1  Work Plan .................................................................................................................43 

List of Figures

Figure 1-1  Limpopo River Basin Showing Location and Riparian States...................................... 2 Figure 1-2  Cyclone Vulnerability Map of Eastern Southern Africa ............................................... 5 Figure 1-3  Inundated Area along Limpopo River during February 2000 Floods ............................ 6 Figure 1-4  Botswana Telemetry Network and Village Flood Watch ............................................. 7 Figure 2-1  Limpopo Telemetry Network Implemented by ARA Sul as Planned ...........................25 Figure 2-2  Telemetry System for Limpopo River FFEWS as Constructed ...................................26 Figure 2-3  Model Flood Forecasting and Early Warning System................................................28 Figure 2-4  Three Operational Components of NWSRFS ...........................................................29 Figure 2-5  MIKE FLOOD WATCH Infrastructure Framework .....................................................31 Figure 2-6  MIKE FLOOD WATCH Interfacing with Model Tool ..................................................32 

iii

ACKNOWLEDGEMENT This proposal to improve the Flood Forecasting and Early Warning System for the Limpopo River Basin has been prepared on behalf of the World Meteorological Organization by Mr Orborne N. Shela, Majiatua Engineering Services, Lilongwe, Malawi with cooperation and support from the Limpopo Water Course Secretariat and the riparian states of Botswana, Mozambique, South Africa and Zimbabwe. The World Meteorological Organization appreciates and acknowledges the excellent support provided to Mr Shela by:

• Mr. Sergio Sitoe, Interim Executive Secretary of the Limpopo Water Course Commission Secretariat;

• Mr. Brink Du Plessis from the caretaker/flood management office manager for SADC–HYCOS at Department of Water Affairs, South Africa;

• Mr. Chivambo, Director General of ARA Sul and his staff tor arranging logistics for and conducting the field trip to Lower Limpopo River, and

• Officials in National Hydrological Services, National Meteorological Services, National Disaster Management Services and other stakeholders in the riparian countries for their support and contributions during consultations.

iv

ACRONYMS AND ABBREVIATIONS

ADB African Development Bank ARA Sul Regional Water Administration for the South in Mozambique CSD Circuit Switching Data CHR Commission for the Hydrology of the Rhine River CS Calibration System DHI Danish Hydraulic Institute DFID British Department for International Development DMCPA Disaster Management and Civil Protection Authorities DWA Department of Water Affairs (South Africa) DWAF Depart of Water Affairs and Forestry (South Africa) EPS Ensemble Prediction System ESP Ensemble Streamflow Prediction EU European Union FFEWS Flood Forecasting Early Warning Systems FFG Flash Flood Guidance FMM Flood Management and Mitigation FMMP Flood Management and Mitigation Programme GDPFS Global Data - Processing and Forecasting System GeoSSFM GeoSpatial Streamflow Forecasting Modelling GIS Geo-Information System GIZ Germany’s Deutsche Gesellschaft für Internationale Zusammenarbeit GPRS General Packet Radio Services GSM Global System for Mobile Communications, GTS Global Telecommunication System HEC Hydrologic Engineering Centre of US Army INGC National Institute of Disaster Management of Mozambique IWRM Integrated Water Resources Management LAN Local Area Network LBPT Limpopo Basin Permanent Technical LBPTC Limpopo Basin Permanent Technical Committee LIMCOM Limpopo Watercourse Commission LINUX A Computer operating system developed by Linus Torvalds of Sweden LNB Low noise block-down converter - satellite receiver component MAP Maximum Amount of Precipitation MRC Mekong River Commission NDA National Directorate of Water Affairs of Botswana NGO Non Governmental Organization NHS National Hydrological Service NMC National Meteorological Centre NMHS National Meteorological and Hydrological Service NMS National Meteorological Service NWHS National Weather and Hydrological Service NWP Numerical Weather Prediction NWS National Water Services of United States

v

NWSRFS National Weather Services’ River Flow Forecasting System PC Portable Computer PIU Project Implementation Unit RFEWS River Forecasting and Early Warning System RGS River Gauging Station RMC Regional Meteorological Centre SADC Southern African Development Community SADC-HYCOS SADC-Hydrological Observing System SARFFGS Southern African Region Flash Flood Guidance System SHEF Standard Hydro – Meteorological Exchange Format SIDA Swedish International Development Agency SMS Short Message Services SQL Structured Query Language - for relational database management SWFDP Severe Weather Forecasting Demonstration Project UN United Nations UNDP United Nation Development Programme UNIX multitasking, multi-user computer operating system USAID United States Agency for International Development USGS United States Geological Survey WIGOS WMO Integrated Observation System WIS WMO Information System WMO World Meteorological Organization

vi

EXECUTIVE SUMMARY

1. BACKGROUND This proposal to improve the Flood Forecasting and Early Warning System (FFEWS) for the Limpopo River Basin was prepared for the World Meteorological Organisation for the benefit of Limpopo Watercourse Commission, a trans-boundary organisation established by riparian states (Botswana, Mozambique, South Africa and Zimbabwe). The LIMCOM recognizes the issues associated with integrated water resources management and, in particular, flood risk management in the basin. In this regard it has embarked on the development and implementation of a Strategic Framework and Integrated Water Resources Management (IWRM) Plan with programmes and activities centred on disaster risk reduction and water quality and water allocation management.

The request for the development of this proposal is a contribution to the LIMCOM Strategic Plan implementation and its preparation and adoption is in response to growing concerns of a lack of coordinated and the non-availability of a real time FFEWS in the Limpopo River. This is despite the Basin experiencing and being impacted on by devastating floods. In recent past, the river has experienced major floods in, for example, 2000, 1999, 1996, 1985, 1997, 1975, 1972, and 1967. The worst of these floods occurred in 2000 when:

• More than 500 people died; • More than 2 million were displaced; • The Mozambique’s part of the river swelled from less than 100 metres wide to 10

to 20 km wide for a more than 100 km stretch and inundated more than 1,400 km2 of farm land; and,

• More than 20,000 cattle were drowned in Mozambique alone. These devastating floods in the Limpopo River Basin are often related to the cyclonic activity, as well as other heavy rainfall events. However, the river also experiences droughts and is at times a basin under water stress with some of the worst droughts occurring in 2003, 2002, 1995, 1994, 1992, 1991, 1987, 1984, 1983, 1981 and 1980. These droughts also have a major impact on the peoples of the Limpopo River basin. Attempts have been made in the past to monitor Limpopo River flows and mitigate droughts and flood disasters but have met with limited success. All riparian countries have river gauging and weather stations where some relevant data and information for flood forecasting and early warning system development and operation is collected. This includes that collected from SADC-HYCOS stations. The SADC-HYCOS project, which ran between1998 to 2005 or thereabouts, installed some 28 stations in the basin with at least 5 in each country except in Mozambique where it had two. In addition to this, a USGS assistance project installed 8 automatic real time gauges in Botswana and a World Bank project in Mozambique installed 19 real time gauges in Lower Limpopo (Mozambique). Some flood forecasting and early warning models were also installed in Mozambique using MIKE 11 and its upgrades and GeoSSFM. Staff personnel in the NHS were trained and the capabilities to install and service the telemetry equipment were developed. However, less than 14% of these real time stations are now working and the FFEWS is struggling to be operational, in the main due to inadequate project design, technical issues and maintenance problems.

vii

The vulnerable communities and general public that are impacted on by the flood disasters in Limpopo River Basin are the primary stakeholders that would benefit from a fully operational and effective FFEWS. However, the current levels of capability in this trans-boundary river Basin impose considerable challenges to efforts of improving the Limpopo River Basin FFEWS. The challenges of trans-boundary flood risk management are being somehow lessened by regional cooperation besides LIMCOM. Regional cooperation endeavours include the sharing of meteorological data and information among National Weather Services of the region through World Meteorological Organization’s Global Telecommunication System (GTS). This facilitates forecasting the likelihood of severe weather, including floods, using meteorological data and information. It can also provide the National Hydrological Services the necessary input to flood forecasting. This puts National Weather Services and National Hydrological Services in the forefront of the regional operational stakeholders with responsibilities for the Limpopo River Basin FFEWS. Other operational stakeholders are the National Disaster Management Authorities in the four Limpopo Basin countries. The Limpopo Permanent Technical Committee, which is also an advisory organisation to the four riparian governments, is a strong regional stakeholder. It has task teams under it and the Flood Task Team is one of them, which now also serves under LIMCOM. From a funding perspective, the donors such as the World Bank, DFID, USAID, GIZ and African Development Bank, are providing technical and financial assistance in disaster risk management and trans-boundary water resources management, and this makes them also stakeholders in the Limpopo River Basin FFEWS. SADC acts as project catalyst stakeholder as all the riparian countries are its members and SADC itself is implementing trans-boundary water resources management projects including those on floods. The SADC Water Division’s Revised Protocol on Shared Watercourse sets the policy, legal and institutional environment for establishment of trans-boundary river basin cooperation institutions such as LIMCOM and determines how cooperation in building instruments and tools for trans-boundary river basin management, like FFEWS, should be implemented. It is also important to note that organizations that deal with flood disaster relief, preparedness, mitigation, response and recovery coordination and implementation are also important stakeholders, with defined roles and responsibilities. However, this review and evaluation of the existing FFEWS indicates that the Limpopo River Basin continues to face flood risk management challenges which include:

(i) Inadequate water resources information and flood monitoring systems necessary for collecting data and information. Such data and information are fundamental to establish, operate and maintain an effective and efficient river flow forecasting systems or FFEWS;

(ii) Limited data exchange and technical cooperation between National Weather Services and National Hydrological Services within the country and among National Hydrological Services within the Limpopo Basin. This prevents sharing of hydro-meteorological data for real time flood forecasting and warning;

(iii) Uncoordinated and incomplete forecasting and warning systems. The few systems that do exist cover a small portion of the basin and are not, therefore, representative, accurate, not fully developed. There are no river forecasting systems which would be used as decision support in water allocation, water quality

viii

and pollution control. This further poses a challenge to LIMCOM who are expected to advise riparian countries on disaster, water allocation and water quality pollution control management; and,

(iv) Limited institutional and capacity development. Currently there is no regional centre that can coordinate telemetry work, lead in river flow/flood forecasting and early warning system development and issue regional flood warning products. There are also limited or inadequate resources in the National Hydrological Services (NHS) and Disaster Management authorities and inadequate technical and administrative competence among the existing personnel at NHS’s.

2. REVIEW OF THE FLOOD FORECASTING AND WARNING SYSTEMS The challenges described above are being manifested in emerging opportunities but limited coordinated development regarding flood forecasting and early in the Limpopo River Basin. The existing opportunities in flood forecasting developments in the southern African include:

(i) The Southern African Region Flash Flood Guidance System (SARFFGS, project, which is developing tools for flash floods forecasting and warning using the estimation and forecasting of antecedent rainfall as well as excess rainfall amounts that trigger flooding in catchments of less than 200 km2, using satellite and radar derived estimates of rainfall that are to some extent verified by observed rainfall at weather and rain-gauge stations;

(ii) The Severe Weather Forecasting Demonstration Project for Southern Africa is a project that uses Numerical Weather Prediction capabilities to forecast severe weather events, including the intensity and movement of rainfall events and severe winds. The data and information gathered facilitates the development and use of severe weather forecasting including conditions likely to give rise to floods;

(iii) The SADC-HYCOS project supported the installation of hydrological monitoring stations for real time hydro-meteorological data acquisitions at selected stations throughout SADC under Phase I and II. However, Phase II has yet to be finalised and those stations that have been installed have problems related to ongoing operation and maintenance. Currently, only 20% of the SADC-HYCOS stations in Limpopo River Basin are fully operational. Despite this, it has been recommended that all SADC-HYCOS stations in Botswana, South Africa and Zimbabwe parts of Limpopo River catchment be reactivated, repaired or installed to form part of the Limpopo River Basin FFEWS telemetry network because they represent an adequate network for FFEW for the catchments in the three countries. Perhaps consideration could be given to the repackaging of those elements of the SADC-HYCOS as a Limpopo-HYCOS;

(iv) The South Africa River Flow Flood Forecasting System has a telemetry system that has at least two real time communication mechanisms at each station. This has ensured availability of data at all times in the Orange River basin. The system also made use of a National Weather Service (USA) River Flow Forecasting System (NWSRFS) Model as the basis for a FFEWS in the Orange River. The system was developed with assistance from US government. The assistance included provisions for capacity building. However, the system has not been maintained due to hardware problems and inadequate data being generated by national telemetry stations in South Africa. Currently, the system makes use of a home grown model for flood forecasting based on flood routing and limited

ix

rainfall/run-off modelling due to problems with NWSRFS based FFEWS. At this stage, it seems impractical to extend the NWSRFS model to the Limpopo River Basin as a whole; and,

(v) The ARA Sul Limpopo River Flood Forecasting and Early System which is based on MIKE 11 and was developed in 2005 or thereabouts, following the installation of a radio based telemetry system in the Mozambique’s part of Limpopo River Basin. However, the MIKE 11 based FFEWS is not working due to data transmission issues from the telemetry system and problems related to the boundary conditions of the model. There are also concerns with the routing component as the cross-sections used were more than the channel width. ARA Sul is now using a GeoSSFM based FFEWS with data collected from satellite and radars in South Africa and data from its own telemetry system as input to rainfall/run-off modelling component. This FFEWS can only give indicative results as contributions from Botswana and Zimbabwe are still not incorporated into the FFEWS.

Thus the Limpopo River basin does not have a coordinated and effective basin wide and fully functional FFEWS to take advantage of the above opportunities and those offered by technologies advances in establishing an ideal FFEWS. An ideal FFEWS requires end-to-end components from flood monitoring systems, through flood models to effective warning dissemination mechanisms and connections with emergency and disaster response. These components include:

(i) Real time flood and related data monitoring where data and information on hydro-meteorological conditions are measured and collected by telemetry equipment, radars and satellites and transmitted in real time using telephones, satellite, radios and cellular telephone technology throughout the basin by various stakeholders and easily shared among all stakeholders concerned;

(ii) Flood forecasting models that are calibrated and operated to generate highly accurate and timely forecasts that are used to developing actions intended to save lives and protect property from floods. The prominent models are the NWSRFS and MIKE 11 and its upgrades such as MIKE FLOOD WATCH, with the latter having an edge over the former when it comes to use in a trans-boundary river basins as it allows forecasting from at least two locations, meaning it can be operated from a regional centre and four sub-centres in the four riparian states;

(iii) Flood early warning and response where, subject to the agreement of all countries, a regional centre should not only issue warnings but also advisory statements on what should be done and what should not be done before, during and after a flood warning is in effect as the forecasters have better knowledge of the river behaviour; and,

(iv) Trans-boundary FFEWS management practices which are being used to coordinate and implement FFEWS involving at least two countries. The experiences in Rhine and Mekong River basins are regarded as the best examples. In Rhine River FFEWS is done by individual NHS’s in each country using data and information readily available from that country and other riparian countries as well, which is facilitated by the Commission for the Hydrology of the Rhine established by scientific institutions interested in the hydrology of the Rhine River. In Mekong River basin the FFEWS was implemented after the establishment of Mekong River Commission by lower Mekong riparian states of Cambodia, Laos, Thailand and Vietnam. It (Mekong Commission) has successfully established the FFEWS based on MIKE 11 and its upgrades with a regional centre

x

in Cambodia. The setting up of Mekong River Basin Commission is similar to that in Limpopo River Basin where LIMCOM is in a position to use the riparian agreement to establish an improved Limpopo River Basin Flood Forecasting and Early Warning System.

3 RECOMMENDATIONS FOR IMPROVED LIMPOPO RIVER BASIN FFEWS The limited resources and development in the Limpopo River Basin make it imperative to recommend that an improved FFEWS for the Limpopo River Basin be designed, financed and implemented on the basis of:

• Building upon existing resources and efforts particularly in improving its monitoring network for data and information acquisition systems and existing flood forecasting models;

• Reorientation of its objectives and outputs to that of a river flow and flood forecasting and early warning system; and,

• Institutional development and capacity building to consolidate the capabilities of the basin’s states to operate and maintain the systems.

These three areas of improvement are further summarised as follows. 3.1 IMPROVEMENT OF LIMCOM MONITORING SYSTEMS

It is recommended that detailed design, rehabilitation or installation of the proposed LIMCOM telemetry stations, data collection programmes and respective manual of standards, guidelines and procedures be prepared, commissioned and implemented. The following activities are recommended to achieve this: -

(i) Consolidation and modernizing of the telemetry network based on the SADC-HYCOS stations in Botswana, South Africa and Zimbabwe and the ARA Sul telemetry stations in Mozambique part of the Limpopo River Basin. All the stations should be rehabilitated and installed with appropriate sound and robust equipment decided upon after thorough field surveys and assessment of the existing equipment and conditions of the stations. It is also recommended that all stations should have satellite or mobile phone based communication systems with backup radio or other forms of communications at each station wherever possible but especially at stations critical to modelling;

(ii) Harmonization of data collection and processing routines where different methods, procedures, standards and guidelines are used for data collection. A standard guidance manual should be produced. The operation and maintenance for all of the LIMCOM telemetry stations should use and follow the manual to ensure compatibility and uniform accuracy of data collected from all the stations and used in the FFEWS;

(iii) Development and launching data sharing systems where it is recommended that protocols, procedures and mechanisms for sharing data between and among NHS and NWS within and among riparian states be developed and commissioned early into the project. The communication software such as GTS and WMO Integrated Global Observation System (WIGOS) and WMO Information System (WIS) would be considered to establish such real time sharing mechanisms; and,

xi

(iv) Trans-boundary telemetry network management where it is recommended that the LIMCOM FFEWS telemetry stations operation and maintenance is audited by LIMCOM itself. LIMCOM may also wish to provide the training/expertise hub for the operational and maintenance activities.

3.2 THE RIVER FORECASTING SYSTEM The decision support requirements of LIMCOM in the integrated water resources management for the basin necessitate the reorientation of the objectives for improving the FFEWS to be support and inform decision making regarding:

• Flood forecasting and warning services needed for safety of life and protection of property in vulnerable areas;

• Flood forecasts to safeguard and be used in the operation and maintenance of water resources infrastructures such as dams and reservoirs;

• Flow forecasts to be used to optimize water resources conservation decisions in the operation of dams and reservoirs by assisting in optimised scheduling of reservoir release operations; and,

• Flow forecasts to be used to optimize water allocation and water quality pollution control by rational and reasonable control of water right abstractions from and consent to discharge water wastes into the river and its tributary basins.

It is, therefore, proposed that in improving the Flood Forecasting and Early Warning System for the LIMPOPO River basin, consideration is given to its design, development, operation and maintenance that meet and satisfy the above listed wider objectives. In effect, the resulting River Forecasting System would have the following outputs:-

(i) Regional River Forecasting Centre established and running; (ii) River Forecasting Models fully calibrated and running with the RFEWS as a major

component and providing Flood Forecasting and early Warning capabilities; and (iii) A manual for the updating, operation and maintenance of RFEWS’s.

These outputs are expected to be achieved with activities that include:-

a) Establishment of Regional River Forecasting Centre where LIMCOM will survey, select, negotiate and partner with one of the existing institutions to be the host of the river forecasting centre. Each riparian state will also designate its own institution to serve as sub-centre for the Limpopo RFEWS;

b) Development and commissioning of the river forecasting models complete with flood forecasting and early warning system based on, for example, MIKE FLOOD WATCH; and,

c) Development of the Manual of Standards and Guidelines of the Limpopo RFEWS, which should be used at the regional centre and sub-centres in each riparian states and will cover updating, operation and maintenance of RFEWS and what to do and not to do before, during and after flood warnings including activities during flood watch, flood disaster relief operations and recovery period and crafting warning and advisory statements.

xii

3.3 RFEWS INSTITUTIONAL DEVELOPMENT AND CAPACITY BUILDING The institutional development and capacity building programme development and implementation will also commence with survey and evaluation of the capacity of existing NHS in terms of resources and personnel required to design, install, update, operate and maintain the telemetry stations and the river forecasting and early warning system as described above. Thus the RFEWS institutional development and capacity building will be accomplished by the following activities:-

(i) Technical Capacity Building Programme development which will involve surveys and evaluations of the institutional capacity and preparation of respective RFEWS institutional development and capacity building programme; and

(ii) Capacity Building and training where the regional centre and national sub-centres of the RFEWS should be resourced and the staff trained through conducting the adopted RFEWS institutional development and capacity building programme and using the developed manuals, as among the training materials.

The development and implementation of the RFEWS institutional development and capacity building programme should take lessons from the development and implementation of the SADC-HYCOS. The emphasis should, therefore, be on building capacity for updating, operation and maintenance of the entire and all components of river flow forecasting system. 3.4 INPUTS AND COST ESTIMATE FOR THE DEVELOPMENT OF RFEWS As there is some existing infrastructure, it is estimated that the work of establishing the RFEWS will require a period of two and half years with the main activities lasting for eighteen months. It is also assumed that the regional centre and national sub-centres will be housed in the existing agencies and buildings. The majority of the telemetry equipment installation works can and will be carried out by the trained staff in each country.

The technical assistance envisaged is that that of a Project Manager and Engineer, who will comprise the core Project Team. The Team will update this report and procure extra equipment, install all the equipment with assistance from riparian trained staff and carry out capacity building and training. The team will also assist in procurement of the services of flood forecasting modelling consultants/contractors who will supply, install, calibrate and commission the RFEWS and train counterpart staff accordingly. This may take a little longer than 18 months but it is expected that the personnel that LIMCOM will hire locally will follow up with the contract until the RFEWS is handed over. It is proposed, therefore, that only specifically identified input requirements such as RFEWS hardware and software; and, provision of technical assistance will comprise the project financial input. In this regard, the cost for developing an improved River Flow and Early Warning System for the Limpopo River Basin is estimated to be approximately US$3.0M. The breakdown of the budget and cost estimate of the input can be summarised as follows:-

(i) Improvement of telemetry system and data collection; 400, 000 (ii) RFEWS development contract 1, 500, 000

xiii

a. PC work stations and accessories b. PC servers – one for each RFEWS Centre c. Modelling Software and licenses purchase (4 sets) d. 36 man-month consulting services @ $22,000

(iii) Project Manager & Project Engineer 500, 000 (@ $15,000 & $12,000/mm for 18 months)

(iv) Support Services and Running Costs 600, 000 a. Project Vehicles – two d/c pick-ups @ $60,000 each b. Vehicle running costs at $9,500 per month field work c. Office Running Cost -$ 8,000 per month for 18 months d. Workshops and Training transport and accommodation, etc.

1

LIMPOPO RIVER BASIN

A PROPOSAL TO IMPROVE THE FLOOD FORECASTING AND EARLY WARNING SYSTEM

Chapter 1 : BACKGROUND This is a project proposal for upgrading the Flood Forecasting and Early Warning System for Limpopo River Basin. It has been prepared by the World Meteorological Organization for the benefit of the Limpopo Watercourse Commission (LIMCOM), an advisory and coordinating body established under an agreement by the four riparian states of Botswana, Mozambique, South Africa and Zimbabwe. The objective of LIMCOM is to “...advice ....and provide recommendations on the uses of Limpopo, its tributaries and waters for purposes and measures of protecting, conserving and management of the Limpopo.” LIMCOM has its interim secretariat in Maputo and is the main institution that implements the LIMCOM agreement adopted in 2005. In an effort to achieve LIMCOM’s objective, the secretariat is currently coordinating the preparation of an IWRM Strategic Framework Plan whose objective is to “…develop the capacities (individual, organizational and institutional) in the riparian states for the sustainable management and development of the Limpopo River Basin”. The Plan has an IWRM Programme that has three main themes of:

(i) Disaster management where LIMCOM is coordinating the preparation development and implementation of: a. disaster preparedness, b. early warning systems for floods and droughts; and, c. development and management of water resources infrastructure for

mitigating the impacts of floods and droughts; (ii) Water Quality Management where LIMCOM has activities that include to:

a. promote adoption of common water quality standards for abating trans-boundary water pollution;

b. facilitate development of a trans-boundary water quality monitoring and reporting system; and,

c. Coordinate the implementation of best practices in water pollution control pilot schemes and assessment and dissemination of best practices on abatement of trans-boundary water pollution.

(iii) Water Allocation where LIMCOM develops and implements strategies and action plans that: a. Promote water resources benefit sharing through equitable and

reasonable utilization of water resources in Limpopo watercourse; b. Facilitate water resources and water use monitoring through

dissemination of data and information on water resources and water utilization; and,

c. Promote water availability and efficient use through encouraging water conservation methods that increases water efficiency and water availability in Limpopo Watercourse or River Basin.

LIMCOM itself was established by and currently reports to Limpopo Basin Permanent Technical Committee, LBPTC. The LBPTC was formed in 1998 with the responsibility of

2

advising the four governments on the management of the Limpopo River, including flood management. The four riparian states recognize the importance of cooperation in dealing with the nature of droughts and flood disasters in Limpopo River Basin. Hence, this is a joint proposal for improving flood forecasting and early warning system in the Limpopo River. The recognition arises from the complexity and challenges of basin itself and its hydrology, which are further elaborated in the following sections.

1.1 LIMPOPO RIVER AND FLOODS The Limpopo River is some 408,000 km2 and flows for a distance of 1,750 km from its headwaters near the border between South Africa and Botswana. Its catchment area distribution among South Africa, Botswana, Zimbabwe and Mozambique is 45%, 20%, 15% and 20%, respectively. The Limpopo flows between the border of South Africa and Botswana, then between South Africa and Zimbabwe before it flows through Mozambique. It eventually enters the Indian Ocean at Zongoene near Xai-xai, Mozambique. Figure 1-1 Limpopo River Basin Showing Location and Riparian States

Limpopo River Basin Map – After http://www.limpoporak.com/ The floods and droughts in the basin are further complicated by the disparities in climatic and rainfall distribution with most of its catchment area under semi-arid conditions. However, the

3

catchment is occasionally influenced by tropical cyclones that can dump significant amounts of rain causing phenomenal floods in Limpopo River. Thus it is highly prone to floods and droughts disasters. Table 1.1 shows some of these devastating droughts and floods. Table 1-1 Some Catastrophic Floods and Droughts in Southern Africa

Year

Type of Disaster

Influenced By Cyclone

Affected Areas and Some More Details

2008 Flood Jokwe Zambezi, Púngue, Búzi and Save rivers in Mozambique flooded. Zambezi River flooded for more than two weeks with 258,000 people affected and more than 100,000 displaced in Mozambique

2007 Favio Zambezi River flooded –more than 120,000 displaced and 250,000 people affected in Mozambique. (http://www.care.org/newsroom/articles/2007/02/20070223_mozambique_cyclone.asp)

2003 Floods Delfina Zambezi River flooded. Seven people died and more than 30,000 people displaced in Malawi and more than 400 homes washed away in Mozambique http://www.ncdc.noaa.gov/sotc/hazards/2003/jan

2002-

2003

Drought Drought period for most of rivers on the southern east coast of Africa with some parts of Limpopo Basin affected. More than 43 districts affected in Mozambique, including those in Limpopo River Basin

2001 Flood Dera Zambezi River flooded, with 115 deaths reported and more than 500,000 affected in Mozambique and 340,000 people affected in Malawi. (http://www.irinnews.org/report.aspx?reportid=18976)

2000 Flood Elaine, Gloria and Huda

Limpopo, Maputo,Umbeluzi, Incomati, Buzi and Save rivers severely flooded. Some 640 deaths recorded and more than 2 million people affected, EN1 main road in Mozambique closed for several weeks,

1999 Flood Floods in Limpopo (http://reliefweb.int/node/47853), Pungwe, Buzi and Save rivers and in Inhambane province Mozambique with EN1 closed for several weeks; 100 deaths;70,000 people affected in Mozambique alone

1997 Flood Lisette Floods in Buzi, Pungue and Zambezi rivers; no road traffic to Zimbabwe for weeks; 78 deaths; 300,000 people affected in Mozambique alone

1996 Flood Bonita Floods on all southern rivers of the country; (including Limpopo River) – 200,000 people affected in Mozambique alone

1994-

1995

Drought Southern east of Africa river basins, including Limpopo, with more than 1.5 million in Mozambique alone. Major crop failure and outbreak of cholera epidemic

1991-

1992

Drought Extensive drought in southern Africa countries and some 1.32 million people severely affected

1987 Drought 8000 affected in Inhambane province 1985 Flood Floods in southern provinces of Mozambique including Limpopo River; and

recorded worst flooding in 50 years 0.5 million affected 1983

-1984

Drought Most of the Mozambique affected. Cholera epidemic and many deaths from drought, which further worsened the suffering of the people from civil war war

1981-

1983

Drought 2.46 million people affected in south and central parts of Mozambique

1981 Flood Floods on Limpopo river ;0.5 million people affected

1980 Drought Southern and central parts of Mozambique affected

1.1.1 Hydrology and Floods of Limpopo River The hydrology of Limpopo River Basin is influenced by the highly seasonal distribution of rainfall over the catchment area. 95% of rainfall falls between October and April with peak mean monthly totals in February. The distribution varies quite considerably from as low as a mean of 200 mm per annum in the western most semi arid part of the catchment to a mean

4

of over 1,500 mm per annum in the south middle part of the catchment while the eastern part, near the Ocean, it averages 600 mm per annum. The basin mean annual rainfall is some 530 mm. However, most rainfall events are highly episodic but intense, usually associated with convective thunderstorms and sometimes cyclones. The Limpopo River Basin climatic conditions have led to its hydrology being characterized by flash flows in the headwaters and highly seasonal flows with most streams and a good part of the main channel having a dry river bed during dry season months. The mean annual hydrograph indicates that the mean dry season flows are as low as 20 m3/s (in September and October), and, higher than 590 m3/s at their peak in February. The flood hydrographs in the flood plains at Chokwe, however, show the main river rising rapidly to exceptionally high annual peaks averaging about 1,600 m3/s with severe floods reaching over 17, 750 m3/s while mean peak flood water levels rises from 0.5 m to 5 metres. During severe floods the water levels rise to over 13 metres, for example like those of 2000 floods. Note that the flood alert or warning level at Chokwe is 4.0 metres. The floods in Limpopo River Basin are caused by heavy rainfall from tropical depression formation as well as cyclone induced rainfall in the catchment area (see Figure 1.2). The passage of cyclones in the catchment area is by far the major cause of heavy rainfall resulting in phenomenal floods, shown in Table 1.1. The impacts of these cyclones on flood disasters were significant in 2000, 1999, 1996, 1985 and 19981 as, shown in Table 1.1. Besides, information available further indicates that Limpopo River also experienced major floods in 1955, 1967, 1972, 1975 and 1977. The majority of these floods were associated with the presence of tropical cyclones in or within the vicinity of the Limpopo River Basin. The Limpopo River Basin can be in the path, impacted directly by cyclones or within their vicinity, (as shown in Figure 1.2) as cyclones can cover an area 150 to 1,000 km wide. The worst of these cyclones were those that occurred in February 2000. The first was cyclone Eline, which caused heavy rainfalls throughout the Limpopo River Basin. In Botswana part of the catchment, for example, rain gauges registered more than 1,000 mm in a single storm event, which was more than half of the average annual rainfall total. The exceptional storms resulted in floods occurring throughout the basin in Botswana, South Africa, Zimbabwe and Mozambique. In Mozambique’s flood plain, the Limpopo River itself swelled from less than 100 metres to10 to 20 km wide, as shown in Figure 1.3, and inundated more than 1,400 km2 of farm land and drowned more than 20,000 cattle, apart from the impacts shown in Table 1-1. It is important to note that droughts also impact on the basin as shown in Table 1-1. Droughts further reduce the river flows and water availability significantly, with cases of water scarcity rampant amid increased demand in the basin. This creates high competition for water abstraction as well as demand for waste water disposal, resulting in conflicts of interest among users that can be understood better and sustainably solved with the assistance of analysis and evaluation of relevant data and information of river flows. Management of the water demand, therefore, requires the river flow monitoring and relevant information management that facilitates sound decisions for appropriate sharing of the limited water resources among users and the riparian states.

5

Figure 1-2 Cyclone Vulnerability Map of Eastern Southern Africa

NB. Cyclone Path or Tracks Map (modified after Atlas of Disaster Preparedness and Response in the Zambezi Basin)

1.1.2 Floods Monitoring and Information Management The availability and use of real-time data and information on rainfall, water levels in rivers and their incorporation in flood forecasting models can facilitate forecasting river flows and floods, which can accordingly facilitate issuing of warnings to vulnerable communities and general public against impending floods. There are 2,700 rainfall stations and some 700 river gauging stations (shown in Table 1-2) that provide data and information on river flows/floods in the Limpopo River Basin. The monitored and collected data and information from these stations is processed, analysed, stored and disseminated by the National Hydrological and Meteorological Services or their agents in all the four riparian countries. The National Weather Services use the Global Telecommunication System, GTS, to transmit and share data and information, including that of rainfall, among fellow NWSs in the riparian countries.

6

Figure 1-3 Inundated Area along Limpopo River during February 2000 Floods

February 2000 Limpopo Floods showing Flooded Area (after Limpopo River Awareness Kit) http://www.limpoporak.com/en/river/hydrology/hydrology+of+the+limpopo/flooding.aspx?print=1

However, the majority of these stations collect daily (recording time interval) data and information, which are compiled and archived monthly. The comment column in Table 1.2 indicates a few hydro-meteorological stations that are telemetry or real time reporting or automatic stations that collect data and transmit data to databases or servers or whenever programmed (less than daily time intervals). There are also automatic weather stations that report or disseminate rainfall data daily to National Weather Services and then other regional national weather services, through the GTS. Rainfall estimates can also be obtained from available weather radars in Xai-xai (under rehabilitation) and Pretoria. These are also capable of monitoring and reporting rainfall in real time or near real time. There are 102 and 18 telemetry stations in the South Africa and Mozambique part of the basin, respectively. In July 2011, all the stations were working and reporting in South Africa while in Mozambique only 12 were working although none was reporting to the base station in Maputo. Vandalism, particularly of solar panels and batteries, and a lack of maintenance on damaged equipment, were cited as the main problems affecting the stations that were not working. The damaged (failure to maintain) repeaters at Mapai, Vouga, Mpuza or Combomune are the main reasons for the failure to transmit data to the Chokwe, Massingire and Maputo database stations (for further information read section 2.1.5). There are no real time RGS reporting in Botswana and Zimbabwe. In Botswana a telemetry system was developed in early 2000s with the assistance of United States Geological Survey, USGS. The system was established together with Village Flood watch that utilised the data collected from the telemetry system. Figure 1.3 shows the map with locations of automatic meteorological (reporting barometric pressure, temperature, wind speed, relative

7

humidity and rainfall) and hydrological (water level/flow) stations developed in Botswana in 2001or thereabouts. However, due to vandalism and lack of capacity to operate and maintain, the telemetry system is not working now. In the Zimbabwe part there are no real time or near real time data and information monitoring stations in Zimbabwe part of the basin. The Limpopo River Basin, therefore, has limited telemetry system and the available data and information is inadequate for flood forecasting and early morning development, operation and maintenance. Table 1-2 National Hydro-meteorological Network in Limpopo River

Country Rainfall Stations

River Gauging Stations

Comment

Botswana 212 44 No real time RGS reporting or archiving Mozambique 98 18 19 Real time Reporting RGS South Africa 2,370 545 102 Real time reporting RGS, including SADC-HYCOS stns Zimbabwe ~70 85 No real time RGS reporting or archiving Figure 1-4 Botswana Telemetry Network and Village Flood Watch

Map with locations of the installed and upgraded hydrologic and meteorological stations with telemetry systems (After D.P Turnipseed at http://www.wrri.msstate.edu/pdf/Turnipseed03B.pdf )

8

1.2 FLOODS AND DISASTER RISK MANAGEMENT COOPERATION

1.2.1 Food Disaster Stakeholder Analysis

Floods and disaster management in the Limpopo River Basin impact on:- 1. The vulnerable communities and general public living or having business in the

flood prone areas; 2. The Government ministries and Departments responsible for monitoring and

issuing flood warnings and taking responsibilities in the development and implementing flood preparedness, mitigation, response and recovery; and,

3. The donor community and NGOs that assist in the development and implementation of flood preparedness, mitigation, response and recovery programmes.

The vulnerable communities and general public are mainly in rural areas and the poor living below UNDP poverty line of US $1.25 per day with obviously very limited capacity to be resilient against flood disasters. Flood disasters worsen their poverty as the little they have is often washed away, damaged or lost forever. The general public are the travellers, investors, who travel or carry out work in flood prone areas. They get inconvenienced together with vulnerable communities when roads, bridges, telecommunication, farms, buildings or homes are impacted by floods in one way or another. The LBPTC, LIMCOM and its interim secretariat, the government ministries and departments are the public sector stakeholders. These public sector stakeholders are listed and described for each country are in Table 1.3, including their interest or responsibility. Table 1-3 Public Sector Stakeholders of the Limpopo FFEWS

Country

Department/Ministry

Interest/responsibilities

Botswana Department of Water

Affairs

Policy and implementation on flood monitoring, forecasting and warning Water resources management

Meteorological Services

Policy and implementation of climate and weather data collection and forecasting

Mozambique

National Directorate of Water, DNA

Policy and implementation on flood monitoring, forecasting and warning Water resources management

ARA, Sul Planning, development and implementation of FFEWS and water resources management

National Institute of Disaster Management, INGC.

Policy and coordination of flood preparedness, mitigation, response and recovery

National Roads Authority

Safety of roads and bridges from floods

9

South Africa Department of Water Affairs

Policy development and implementation on flood monitoring , forecasting and warning; and, Water resources management

South African Weather Services

Policy and implementation of climate and weather data collection and forecasting

National Disaster Management Authority

Policy and coordination of flood preparedness, mitigation, response and recovery

Zimbabwe National Weather Service

Responsible for climate and weather data and information

Department of Civil Defence Office

Coordinate flood disaster preparedness, mitigation, response and recovery

Zimbabwe National Water Authority

Planning, development and implementation of FFEWS and water resources management.

At the regional scale, SADC is another multilateral stakeholder as all riparian states are its members. SADC has the Revised Protocol on Shared Watercourses in Southern African Development Community as its legal instrument for cooperation among basin member states in the joint management of trans-boundary rivers, including the basin-wide flood forecasting and early warning system establishment, operation and maintenance. The importance of SADC is the provisions of the Revised Protocol that obligate member states to share data and information such as that required in and generated by the flood forecasting and early warning system. SADC is also running various regional projects on water resources management particularly trans-boundary water resources studies in Zambezi, Limpopo and Orange River basins. The donor community, including the World Bank and United Nation Development Programme, UNDP, has regional and country programmes for disaster risk management. The German International Development Agency, GIZ, also has regional programmes that support trans-boundary water resources management activities such as the work proposed herein. There are also a number of multilateral development agencies such as DIFD, SIDA, EU, WMO and USAID who support various disaster risk management programmes in Southern Africa, by providing technical and financial assistance. The donor community stakeholders are mainly contributing to the implementation of the UN Hyogo Framework of Action for Disaster Risk Reduction and assisting in the creation of a supportive regional environment for flood disaster risk management cooperation in Southern Africa.

1.2.2 Regional Policy, Legal and Institutional Arrangements

This proposal for the improvement of Limpopo River Basin Flood Forecasting and Early Warning System is being made under the auspices of LIMCOM, which has been established under an agreement signed by all the four riparian states. LIMCOM itself is modelled along the lines of the SADC Revised protocol on shared watercourses and its functions, as quoted from the agreement, promote, among others:-

10

1. The equitable and reasonable utilisation of the Limpopo to support sustainable development in the territory of each Contracting Party and the harmonisation of their policies related thereto;

2. The extent and mode in which the inhabitants in the territory of each of the Contracting Parties concerned should participate in the planning, utilisation, sustainable development, protection and conservation of the Limpopo and the possible impact on social and cultural heritage matters;

3. All aspects related to the efficient and effective collection, procession and dissemination of data and information with regard to the entire Limpopo;

4. Contingency planning and measures for preventing and responding to harmful conditions whether resulting from natural causes such as floods and droughts or human conduct as well as emergency situations that result suddenly like industrial accidents;

5. Measures for arriving at settlement of a dispute; and 6. Any other matters affecting the implementation of the Revised Protocol on

shared watercourses in SADC.

Thus LIMCOM has been established in accordance to the Revised Protocol for Shared Watercourses in SADC. It must be reiterated that this is the main policy and legal instrument governing cooperation and integration in the development, management and utilisation of shared or trans-boundary water resources or river basins in SADC.

The other influential regional stakeholder is the Limpopo Basin Permanent Technical Committee, LBPTC, which was established in 1986. It has the mandate of advising member states on different aspects associated with droughts, floods, pollution, water resources planning and development across the basin. It continues to be the main institution that sets policy and initiates legal and institutional developments for joint development and utilization of Limpopo River Basin, including decisions of LIMCOM. The importance of LBPTC can be seen in its achievements that include:

• Established the Flood Task Team in the late 1990, responsible for coordinating

and advising on floods in Limpopo River. The Flood Task team comprise of officials from departments of water affairs and other key stakeholders in the management of flood risk in each of the four LIMCOM states.

• Established the Interim LIMCOM Secretariat in 2008, located in Maputo under the auspices of the National Directorate of Water Affairs in Mozambique (DNA);

• Established the Legal Task Team that advise the LBPTC on legal aspects, such as provisional entry into force of the LIMCOM Agreement, development of the host agreement for the LIMCOM Secretariat and other legal issues related to day to day tasks of the LBPTC;

• Facilitated the ratification process of the Establishment of Limpopo Watercourse Commission Agreement for all member states;

• Developed the Limpopo River Awareness Kit for the Limpopo River Basin from 2009 to 2011, which is serving as the information and knowledge management hub for LIMCOM;

• Established the official www.LIMCOM.org website in 2010; • Has been an effective vehicle for notification on ongoing and future

development projects within the basin and other Water Resources aspects of common interest among the member states; and,

11

• Developed internally the Strategic Framework and IWRM Plan 2011-2015 for implementing the LIMCOM Agreement.

As part of implementing the LIMCOM agreement, LBPTC has appointed the Flood Task Team and Legal Task Team to continue to serve under LIMCOM. LBPTC has also rolled out the Strategic Framework and IWRM Plan 2011 -2015. The Plan sets the vision of LIMCOM as “water security for improved livelihoods in the Limpopo River Basin to meet the needs of the current and future generations”. The mission statement for LIMCOM is proposed as “...to advice riparian states on the governance, management and development of water resources in the Limpopo River Basin through integrated water resources management to improve social equity, promote economic efficiency and ensure sustainable development.” Its goal is to “...develop the capacities (individual, organizational and institutional) in the riparian states for the sustainable management and development of the Limpopo River Basin.” The Plan intends to achieve this through three strategic objectives of disaster management, water quality and water allocation as described in the second paragraph under Background.

1.2.3 Flood Disaster Relief Operations and Regional Cooperation LIMCOM values flood disaster risk management, and aspires to have an end to end flood forecasting and early warning system. All four riparian countries have established disaster management institutions that develop appropriate disaster risk management strategies, programmes and projects, which are implemented by line ministries and national, provincial and local civil defence committees. Flood disaster risk preparedness, response, mitigation and recovery programmes and activities are being coordinated and implemented through these committees. In Botswana this is being done by National Disaster Management Office while in South Africa the National Disaster Management Authority coordinates and deals with disaster risk management. The National Institute of Disaster Management, INGC, and Department of Civil Protection deal with disaster risk management in Mozambique and Zimbabwe, respectively.

At the regional level, LIMCOM through its Flood Tasks Team has continued coordinating communication and exchange of data among Member States in connection to flooding events, including issuing of flood warnings and coordination of regional response. The Flood Task Team was very active and instrumental in coordinating communication of causal flood forecasts, warnings and response among the countries during the February 2000 floods in Limpopo River. However, a regional flood management centre does not exist in the basin and the Flood Task Team members work from their respective offices in the country of their origin. This lack of a real time basin – wide flood forecasting system and other capacity development constraints are the major challenges for setting up and operation of the improved Limpopo River Basin Flood Forecasting and Early Warning System

1.3 Flood Risks Management Challenges The flood risk management challenges are centred on the inadequacies of existing water resources and flood monitoring systems, limited data exchange and technical cooperation; uncoordinated and incomplete flood forecasting and warning systems and limited institutional and capacity development. The following sections elaborate on these challenges that face efforts of improving flood forecasting and early warning system.

12

1.3.1 Inadequate Water Resources and Flood Monitoring Systems There are challenges of improving the inadequate water resources and flood monitoring systems that include:

• While the hydrometric and meteorological monitoring networks in South Africa appear adequate and working effectively in collecting data, there are deficiencies in Botswana, Zimbabwe and Mozambique. Real time or near real time monitoring systems in these countries need improvement. This will enable the collection and transmission of data and information needed for flood forecasting;

• Capacities of using satellite rainfall estimation techniques are emerging but challenges exist particularly in ground proofing with the actual rainfall using estimates from radar measurements or observation gauges. Weather radars at Xai-xai, Gaborone and Bulawayo are needed. The challenges are the resources needed to rehabilitate, install and operate the radars at these sites

• While each country has its own teams for collection, processing and analysis of hydrological data used in flood forecasting, different standards and guidelines are used. The challenge is in how to synthesize the data and information from all the countries and use it in flood forecasting that is expected to be accurate and timely to save lives; and,

• The development and operation of a unified basin wide flood forecasting modelling framework requires adequate historical and current short duration as well as daily data. The challenge is how such models will be developed and implemented with limited availability of the data needed.

The other challenge is the deteriorating hydrological services, particularly in carrying discharge measurements, maintenance and servicing of the gauging stations. This challenge further manifests the challenge of data quality assurance in using data and information from all four countries.

1.3.2 Limited Data Exchange and Technical Cooperation Consolidation of the basin wide flood forecasting and early warning system is a challenge as the existing available real time data and information is not readily available for use. The data in each country is not available and shared in real time. The gauging stations data and information is only readily accessible and available to National Hydrological Services of the country of origin. It should be noted that, the rainfall and other meteorological data are easily or can easily be shared between or among the National Meteorological Services of riparian states, through GTS. However such exchange of data between the riparian states’ National Hydrological Services remains a challenge. In particular, this poses a challenge in the development of a unified modelling system or framework for flood forecasting and early warning.

1.3.3 Uncoordinated and Incomplete Forecasting and Warning Systems The challenges of Limpopo River Basin flood and drought risk management include:

13

• South Africa and Mozambique have some kind of river flow/flood forecasting and

early warning systems individually developed and operated. These are described in section 2.1.1 and 2.1.5. However, they tend to be routing models with limited rainfall-run-off modelling. The models may be good for river flow forecasts during the dry season but not ideal for floods or during rainy season. Besides there is no single unified flood forecasting and early warning system for the entire basin;

• The existing flood forecasting models are partially valid or incomplete as catchment areas outside the countries are either accounted for crudely or not accounted for and the results are estimated within limitations of the data used. The forecasting models, therefore, tell half of the story and are only indicative;

• The absence of a basin-wide river flow forecasting system has limited the extent to which existing and planned reservoirs and dams can be operated to maximise water resources conservation and regulation of the river flows to protect water resources and transport infrastructures from flood destruction; and,

• The limitations of the current decision making tools for water resources conservation and allocation (without a river flow forecasting system for all the year round) is a huge challenge to LIMCOM. The present forecasting and early warning system focuses on flood forecasting and yet functions and responsibilities of the LIMCOM include advising on reasonable and equitable utilisation of water resources and water pollution control.

These challenges need to be taken into consideration in improving the Flood Forecasting and Warning System for Limpopo River Basin.

1.3.4 Limited Institutional and Capacity Development The institutional and capacity development required for improving the FFEWS in Limpopo River Basin has many challenges, which include:

• River flow or flood forecasting efforts being done outside the basin and not coordinated or concentrated at one place within the basin;

• The Interim Secretariat at LIMCOM has limited place and capacity to host the basin wide river flow or flood forecasting and early warning system;

• The National Hydrological Services of the four riparian countries and their Interim Secretariat have limited trained and skilled staff in river flow or flood forecasting;

• The existing computer hardware and software systems at National Hydrological Services and interim LIMCOM Secretariat are inadequate and incapable of handling and manipulating large data and information and require complete flood forecasting models and respective early warnings system; and,

• Unlike the existence of regional weather forecasting centre in South Africa, there is no regional river flow or flood forecasting centre for Limpopo River Basin.

14

1.3.5 Summary of Forecasting Challenges and Conclusions

These requirements and challenges mentioned above are the main reasons supporting the need to improve River Flow (RF) or FFEWS in the Limpopo River Basin. This requires an end to end - user driven system that provides information including effective early flood forecasts and warnings in support of safety of lives, protection of property, the welfare of the general public and the safe operation and maintenance of the infrastructure. The river flow forecasting system is particularly required to ensure LIMCOM has a total decision making tool for providing advice and guidance in support of flood and drought risk management among the member states.

The review of challenges can further be understood in the context of a review of the existing FFEWS capabilities and an ideal flood forecasting and early warning system. This understanding is important in support of the development and presentation of the recommendations in chapter 3. The review is, therefore, presented in the next chapter: Review of Flood Forecasting and Early Warning Systems.

15

Chapter 2 : REVIEW OF FLOOD FORECASTING AND WANRING SYSTEMS

2.1 EXISTING FLOOD FORECASTING DEVELOPMENT IN LIMPOPO RIVER BASIN The existing flood forecasting developments comprise those for estimating and forecasting rainfall and generating real time river flows as well as the actual flood forecasts. The rainfall estimation and forecasting systems include the Southern African Region Flash Flood Guidance system, the Severe Weather Forecasting Demonstration Project for Southern Africa, SADC-HYCOS telemetry system and radars. The actual flood forecasting developments are those involving the South African River Forecasting System and ARA Sul Limpopo Flood Forecasting and Warning System. The review of these forecasting developments in Limpopo River Basin is provided below.

2.1.1 Southern African Region Flash Flood Guidance System

The SARFFGS is a WMO flash flood forecast improvement initiative project for southern Africa and is based at South African Weather Service in Pretoria. Once the guidance is fully developed, the centre will be able to send regional flash flood guidance products to National Weather Services (and National Hydrological Services) in southern Africa, including the riparian states to Limpopo River. The National Weather and Hydrological Services would use these to generate local flash flood forecasts and/or warnings, using SARFFGS. The project includes capacity building and training of expert officials in southern African countries, including forecasters (meteorologists and hydrologists).

A flash flood itself can be defined as a flood causing inundation along flood plains of streams and rivers within a short duration, usually less than 6 hours after the commencement of the storm rainfall and generating a relatively high peak discharge. The flash flood guidance is an advance determination based on the prediction that critical thresholds for certain level of flooding will be surpassed rather than the accurate determination of the magnitude and timing of the flood peak. This is done using inferences of excess rainfall within the storm durations and hydrological conditions that would trigger flooding from such excess rainfall.

The flash flood guidance, therefore, can be defined as an estimate of how much rainfall over a specified time in a small basin is needed to initiate flooding on small streams, usually less than 200 km2. In this regard, the SARFFGS will, once developed fully, assist in or be a system of:

(i) Modelling the flood threat (in particular flash flood) over small basins with • Hydrological information; • Rainfall measured by rain gauges (real time), radar, satellite, etc.; and/or, • Rainfall forecasts;

(ii) Monitoring the hazardous weather conditions where a forecaster: • monitors the weather data from weather/rain gauge stations (real time) ,

radars, satellite images, etc.;

16

• Adjusts satellite rainfall estimates and/or radar rainfall fields based on latest real-time data; and,

• Evaluates flash flood threats based on model outputs and rainfall fields; (iii) Warning information preparation where flash flood advisories and warnings are:

• Crafted based on the identified flash flood threats; and, • Disseminated and issued to vulnerable communities, general public and

Disaster Management Centres. (iv) Response and flash flood preparedness plan implementation where:

• Response plans and routines are activated ; • Protection measures such as closing of bridges are evoked; and, • Areas under flood threat are evacuated.

The SARFFGS products would include calibrated maps or tables of threshold rainfall (Flash Flood Guidance (FFG)) in specific sub-regions or areas. The flash flood occurrence will be forecasted when the maximum amount of precipitation (MAP) in a storm exceeds FFG or there is excess precipitation, EP; the greater the amount of EP the higher the severity of the flash flood. Note that the increase of telemetry stations with capacity to measure and transmit rainfall data in real time will enhance the accuracy and usefulness of the SARFFGS in forecasting and issuing flash flood warnings.

2.1.2. Severe Weather Forecasting System for Southern Africa. The Severe Weather Forecasting Demonstration Project (SWFDP is also a WMO initiative aimed at improving regional and local severe weather forecasting using growing opportunities offered by Numerical Weather Prediction (NWP) System products from NWP centres. NWPs are part of the Global Data-Processing and Forecasting System (GDPFS) for weather forecasting. The SWFDP for Southern Africa, based at South African Weather Service headquarters, is one of several regional centres around the world where demonstration of the use of NWP products in severe weather forecasting is being carried out. It involves use of a cascading (forecasting) modelling approach to provide greater lead-time for severe weather and at the same time contribute to capacity building among the National Weather and Hydrological Services (NMHS) and improving links with National Disaster Management and Civil Protection Authorities (DMCPA) in southern Africa. The goals of the SWFDP are the following:

• to improve the ability of National Meteorological Centres, NMCs, to forecast severe weather events;

• to improve the lead time of alerting of these events; • to improve interaction of NMCs with DMCPA before and during events; • to identify gaps and areas for improvements; and, • to improve the skill of products from GDPFS Centres through feedback from

NMCs.

Conceptually, severe weather forecasting involve one global centre issuing global forecasts, one regional centre issuing regional forecasts from fine tuning the global forecast and a small number of NMHSs located within the area of responsibility of the regional centre to compile and distribute local national forecasts after fine tuning the regional forecasts.

17

The SWFDP project has brought the following benefits to southern Africa, which are expected to be consolidated further after completion in 2012:

• Improvement of the weather forecasting and early warning services in the countries through the enhanced use of modern forecasting and early warning technology such as NWP and Ensemble Prediction Systems (EPS);

• Improved support to disaster risk reduction through the early and timely warning services that contribute to building resilience among vulnerable communities and general public in each participating country;

• Increased support to national forecasters through the guidance products from RSMC forecasters, and additional NWP and EPS output, leading to enhanced confidence of forecasters in issuing forecasts, advisories and warnings.

• Trained forecasters and thus capacity building of NMHSs in using modern forecasting technology such as NWP and EPS;

• Increased number of trained forecasters in southern African countries in skills for modern forecasting information and improved forecasting systems;

• Increased collaboration between forecasters and their local disaster management and news media structures;

• Increased regional coordination between and among NMHSs on carrying out forecasts and crafting and disseminating advisories and warnings;

• Opportunity to share, coordinate, and collate all weather warnings in the region; • Enhanced severe weather warning services for the end-users including the flood

forecasters and the general public; • Improved relationships between NMHSs, Regional Meteorological Centre in

Pretoria, RMC, and Global Centres; and, • Opportunities for southern Africa to contribute in evaluation of the global models’

performance including the usefulness of the products to forecasters (for weather as well as floods).

The SARFFGS discussed in 2.1.1 can also benefit from the SWFDP. It uses products developed and generated from SWFDP and associated NWP and global weather forecasting centres. The SAFFGS has a critical role to play in supplying rainfall data for development, calibration and operation of normal flood forecasting models envisaged in improving the existing Limpopo FFEWS. It will also play a critical role in providing flash flood warning in upper reaches in South Africa, Botswana and Zimbabwe.

2.1.3 SADC-HYCOS Telemetry System The SADC-HYCOS is also part of a WMO initiative which commenced in the mid 1990s aimed at facilitating real time hydro-meteorological data and information collection, processing, archiving and dissemination in the SADC region. It comprises a network of telemetric meteorological and hydrological monitoring stations, a base station in Pretoria where real time data and information is automatically transmitted, processed, archived and disseminated to or accessed by respective member states of SADC. The internet website or direct polling the base station is a way of easily and equitably disseminating or accessing the SADC-HYCOS data and information.

The achievement of SADC-HYCOS in improving real time data acquisition in Limpopo River can be seen in Table 1.3- SADC-HYCOS network in Limpopo River Basin during Phase I and II. It can be seen that Limpopo River Basin has 28 river gauging stations, which

18

comprise 5 in Botswana, 2 in Mozambique, 14 in South Africa and 7in Zimbabwe. These stations have the ability to measure and transmit, in real time, water level, selected water quality parameters (electrical conductivity, temperate, turbidity, etc.) and weather parameters (rainfall, air temperature, relative humidity, etc.). The SADC-HYCOS project has also trained at least two technicians from the four riparian states in installation, programming and operation of the telemetric equipment installed at these stations. They have also been trained in real time water resources monitoring techniques and its electronic data and information management using software called HYDSTRA. The SADC-HYCOS project has also supplied and installed computer hardware and software, including HYDSTRA, at each NHS in the four member states. HYDSTRA is used as the database for data and information downloaded from SADC-HYCOS base station or server in Pretoria. This software is ideal for manipulating real time data and information into input data to flood forecasting models and has GIS interfacing capabilities. The gains in SADC-HYCOS have been counter balanced by widespread failure of the system to provide real time data due to, among others:

• Incompletion of project’s works, where the project supplied equipment for 14 stations, (one in Botswana, 9 in South Africa and 4 in Zimbabwe) ready for installation but were not installed due to lack of funds for technicians’ travel and travelling costs from SADC-HYCOS’s PIU in Pretoria;

• Non-functioning and vandalism of equipment at installed stations, where out of the 14 installed stations in Limpopo River basin, only 4 were working in April 2011 and that the 10 stations out of order are in this status mainly due to vandalism and theft of solar panels and two stations were out of order due to flood damages in Mozambique; and,

• Lack of local operation and maintenance support to rectify operation and maintenance problems in the fields, despite training being provided.

These shortfalls have been there since Phase II was completed in 2010 and to some extent even before this date. The National Hydrological Services, NHS, in respective countries, have also been unable to rehabilitate or maintain the SADC-HYCOS stations. It appears that the capacity building to ensure local support for operation and maintenance of the equipment built in into the design and implementation of the project has not be sufficient or adequately implemented or supported at the national level to ensure sustainability of the operation of the stations. The lessons learnt in SADC-HYCOS project design, implementation and operation phases are relevant input in the designing project proposal for improving and consolidating flood forecasting and early warning system for Limpopo River. Capacity building should include development of local support and resilient telemetry systems (to vandalism, theft and other forms of tampering). The operation and maintenance needs, after the project life, should be built in into the project design. It is also important to consider a PIU becoming the regional centre where LIMCOM should be responsible for its ongoing sustainability.

19

Table 2-1 Limpopo Basin SADC-HYCOS Telemetry Network Station List

Country Stn No. Station Name Station Description Status of SADC-HYCOS Station as of April 2011

68044 Nata Nata Stream @ Nata Out of order

68149 Seleka Farm Limpopo River at Seleka Farm Out of order

68150 Buffel’s Drift Limpopo River @Buffels Drift Damaged by Baboons

68238 Gaborone Dam Notwane River @ Gaborone Dam Operational but faulty probe?

68246 Ramotswa Notwane River @ Ramotswa Out of order due to Vandalism

Botswana 68080 Palapye Losane River @ Palapye Not yet installed

SubTotal 5 Number of Stations working 1

67328 Limpopo Limpopo River @ Combomune Out of Order due to flood damages Mozambique 67329 Pafuri Limpopo River @ Pafuri Out of Order due to flood damages

SubTotal 2 Number of Stations working Nil

A5H006 Limpopo Limpopo River @ Sterkloop Not yet installed

A6H035 Mogalakwena Mogalakwena River @ Leneisrus Not yet installed

A7H008 Limpopo Limpopo River @ Berti Bridge Status not known but not reporting

A7H010 Sand Sand River @ Waterproof Not yet installed

A8H000 Limpopo Limpopo River @ Shashe/Limpopo confluence Not yet installed

A8R001 Nzelele Nzelele River @ Nzelele dam Not yet installed

A9H000 Limpopo Limpopo River @ Pafuri Not yet installed

A9H013 Mutale Mutale River @ mutalebend Not yet installed

B7H007 Olifants Olifants River @ Oxford Not yet installed

B7H015 Olifants Olifants River @ Mamba Status not known but not reporting

B8H008 Great-letaba Leteba River @ Leteba Farm Operational

B8H018 Letaba Leteba River @ Engelhard Dam Not yet installed

B8H034 Great-letaba Leteba River @ Black Haron Dam Operational

South Africa

B9H003 Tshinane Tshinane River @ Kanneidood dam Out of order due to vandalism

SubTotal 14 Number of Stations working 2

20

Country Stn No. Station Name Station Description Status of SADC-HYCOS Station as of April 2011

67993 Nuanetsi Nuanetsi River @ Malapati Bridge? Status not known but not reporting

67988 Mzingwane Mzingwane River @ Doddieeburn Status not known but not reporting

67994 Bubye Brigde Bubye River @ Bubye Bridge Not yet installed

67995 Bubye Chikwarakwara Bubye River @ Bubye Chikwarakwara Not yet installed

67989 Shashe Confluence Shashe River @ Limpopo/Shashe Confluence Not yet installed

67978 Tokwe/Runde Conflue. Tokwe River @ Tokwe-Runde Confluence Not yet installed

Zimbabwe 67777 Aurelia farm Musengezi River @ Aurelia Farm? Operational

SubTotal 7 Number Stations working 1

21

2.1.4 South Africa River Flow Flood Forecasting Systems The Department of Water Affairs in South Africa has a “River System Flood Management” centre that runs an in-house built flood forecasting and early warning management system for the Vaal/Orange River Basin. The main objectives of flood management are:-

• Ensuring the safety of the water resources and related infrastructures by optimally operating reservoirs to give room for and attenuating the peaks of the forecasted floods;

• Minimisation of flood losses and damages through issuing warnings about the forecasted floods; and,

• Ensuring that the dams are 100% full at the end of the flood season by cautiously holding releases from the dams during flooding season.

The forecasting system used is a network of telemetry hydro-meteorological stations (telemetry system) equipped with at least two telemetry systems; base station that retrieves, receives, processes and archives data from telemetry systems; and, flood forecasting models. The telemetry system component comprises GPRS Satellite, Circuit Switching Data (CSD) and SMS based data transmission systems which are elaborated as follows:

• GPRS based data transmission systems. The General Packet Radio Service, GPRS, stations transmit data via cellular phone networks and DWA uses four suppliers for this service (GPRS Data, GPRS ProDesign, 4Water and OTT (HYDRAS3). These services are paid for and are expensive (around US $ 1,700 per month).

• The Satellite data (EUMETSAT METEOSAT 9) system transmits data freely by arrangement with WMO under a licence but require high end PCs due to high volume of data being handled and also requires satellite dish with appropriate low noise block-down-converter (or LNB). SADC-HYCOS uses this type of transmission.

• The CSD data transmission uses OTT (HYDRAS3 and ProDesign) with both systems collected off site. The Short Message Services (SMS) data transmission uses 4Water (CELLO), East Coast and UBUNTO systems.

The DWA has specialized personal computers, PCs, and software that receive, process and manipulate the data into Standard Hydro-meteorological Exchange Format (SHEF) or usable and acceptable data format for input into flood forecasting models. The database PC is then populated with SHEF data, which is put on LAN, HYDSTRA and internet by a data distribution PC. This distributed data includes that from SADC-HYCOS stations. It is also this data used in flood forecasting or disseminated to National Disaster Management Centre. The data includes river flow data, already processed from stage after using respective stage ratings, accordingly, whenever such information is available.

The DWA uses in-house developed simple rainfall/runoff models to import real time data from HYDSTRA and generate hydrographs. These hydrographs are routed through reaches and added each time the reaches meet at a confluence. The model uses determined travel times to route hydrographs between confluences, which keep on adding rainfall/run-off modelled hydrographs of in between streams and local run-off within each reach. The

22

resulting river flow/flood forecast data and information is used to make decisions for operating dams to conserve water or discharge more water from reservoirs to protect their dams. It is also used to generate a variety of flood forecast products such as flood hydrographs, flood inundation maps and flood warnings. The forecast products are also used to optimise hydropower generation throughout the Vaal/Orange River system.

The DWA, with assistance from USA National Weather Service, also developed a Flood Forecasting System for the Vaal/Orange River based on the National Weather Service River Flow Forecasting System (NWSRFS), running on a LINUX operating system. This was in response to DWA request for assistance to develop a more reliable river flow forecasting system for South Africa. The project was executed in three phases, as follows:

(i) In the first phase of the project, DWA was provided with computer and was assisted in installing a functional prototype river forecasting system for the Vaal River for advance prediction of inflows upstream of the Vaal Dam. The prototype demonstrated the potential uses of an interactive forecasting system in predicting reservoir inflows in South Africa.

(ii) The prototype model parameters were updated and additional training on the system setup and operation was conducted during the second phase of the project. This updating of model parameters upgraded the system which became more useful as it produced more realistic results.

(iii) The third and final phase comprised DWAF personnel having a three week intensive hands on training at NWS Riverside’s Fort Collins, Colorado office, on the NWSRFS data analysis, model calibration and system operations. During the training period, the DWAF NWSRFS forecasting system was expanded to include the Middle and Lower Vaal River sub-basins, and the portion of the Orange River upstream of their confluence with the Vaal. The DWA and NWS staff also fine-tuned model parameters for the Upper Vaal, and established regional parameter sets for the remaining sub basins. The third phase ended with NWS Riverside Office personnel completing the initialization on the Vaal-Orange system, implemented Ensemble Streamflow Prediction (ESP) on the entire system, and provided additional advanced NWSRFS training to DWA staff in South Africa.

The NWSRFS based Vaal/Orange River Forecasting System is a large and data hungry model requiring huge data storage space and high speed computers. It therefore require the use of special PCs. The PC hosting the NWSRFS system started having problems some years after 2004 and the NWSRFS Vaal forecasting system has not been used since. NWSRFS is further discussed in section 2.2.2. It is likely that the system can be used for Limpopo River considering the amount and quality of data that would be generated by revamped SADC-HYCOS stations, SWFDP and SAFFGS. However, it cannot be adapted to Limpopo River Basin at the moment unless a new relevant PC is secured and the system re-installed and checked for its validity and robustness.

2.1.5 ARA Sul Limpopo Flood Flow Flood Forecasting System The Regional Water Administration in the South, ARA Sul, is responsible for water resources management in the southern part of Mozambique, including the area under Limpopo River basin. It is, therefore, responsible for the Limpopo FFEWS, which comprise: -

23

• the telemetry network with 19 stations installed by itself with base stations capable of receiving data from all stations at Massingir, Xai Xai and ARA Sul Headquarters in Maputo;

• two stations out of the above were also equipped with SADC-HYCOS instruments but damaged by floods;

• the MIKE 11 and USGS Geo-spatial flood forecasting models; and, • three forecasting centres at Massingir Dam, Macarretane weir (which has been

moved from Xai-Xai together with the telemetry base station) and ARA Sul Headquarters in Maputo.

Its telemetry network is shown in Figure 2.1 and has nine green boxes representing proposed rain gauge network (not yet installed), 4 ocean blue boxes being water level stations only 10 red circles being stations with water level and rain gauge network, the 3 black circles being the forecasting centres or base stations and the 10 orange circles being radio repeaters. The status of the network has already been discussed in 1.1.2 and further amplified in Table 2.2 while Figure 2.2 shows the network as constructed.

Table 2-2 Limpopo FFEWS Telemetry Network in Mozambique

Station Number & Name

River/Rain Gauge Status Comments

L01 – Pafuri Limpopo River O, R, WL Em risco de erosao e Sem Guarda. L02 – Monte Bipi Muenezi River O, R, WL

L03 – Vouga Singuisi River NI, Sem Guarda L04 – Mapai Limpopo River Operational Ameacada de Erosao

L05 Not installed L06 Not installed

L07 Not installed L08 Not installed

L09 Not installed L10 – Parque IO, R Sem paineis solares (roubo).

L11 – Parque IO, R L12 – Massingir Olifant River Operational

L13 – Combomune Limpopo River Operational L14 – Mabalane Limpopo River O, R, WL Ameacada de Erosao.

L15 – Confluencia Limpopo River Operational L16 – Macarretane Limpopo River O, R, WL Pluviometro perdido devido a ventos fortes.

L17 – Chokwe Limpopo River O, R, WL l Erosao na Caizxa de Sensor. L18 – Mohambe Limpopo River IO, R, WL al Queda da Toree.

L19 – Sicacate Limpopo River IO, R, WL Vandalizado cabo de sensor Fotoelectrico. L20 – Xai-Xai Limpopo River O, R, WL Sincronizada com a Respectiva Central.

L21 – Zongoene Limpopo River O, R, WL Corrosao da Torre e Levantamento das sapatas da Torre

L22 – Chibuto Changane IO, WL Sem Bateria L23 – Maqueze Changane NI, R

Radio Repeators Monte Bipi Operational

Mapai Inoperational Vouga Inoperational

Mpuza Inoperational Novos paineis nao repostos devido a falta de segurance

Combomune Inoperational Queda dos suportes dos cabos e danos no Source: Modified from personal Communication with Officials of ARA Sul

24

The discharge measuring stations in Limpopo River are mostly located immediately downstream of the bridges except at Macarratane. The rating curves are usually unstable and require high rate of maintenance with frequent and all round range discharge measurements. This is expensive and can be deceiving as, during floods, rating curves may change due to channel scouring or sedimentation taking place.

Besides, most stations are out of order and not functioning due to:

• vandalism with batteries and solar panels missing; • radio masts damaged from vandalism as well as corrosion (from salt blown from

Indian Ocean), which eats away steel masts hoisting the antennas; • Not yet installed due to lack of equipment; • Radio repeaters damaged or having operational problems worsened by

vandalism, corrosion and radio wave interferences such those experienced in Maputo where the base station has been non-operational since installed due to weak signals and wave interference; and,

• Lack of capacity to repair and fix field damages and wearing out of the telemetry system.

This has resulted in ARA Sul having inadequate readily available real time hydro-meteorological data and information from its network. However, ARA Sul has had access to rainfall estimates from SAFFGS or other centres that process precipitation estimates from satellite imagery. However, this can also prove inadequate as verification requires ground proofing from rain gauges on the ground. This has only been possible from Olifant River in South Africa and not that part of catchment in Zimbabwe and Botswana. This and the non installation completion of stations particularly in Changane sub-basin in Mozambique have forced ARA Sul to seek alternative models in forecasting floods in the Basin.

At first ARA Sul installed MIKE 11 “Flood Watch” as the flood forecasting software in 2005 or thereabouts, with flood forecasting centres at Massingir Dam, Xaixai and in Maputo where all FFEWS under ARA Sul are controlled. The MIKE 11 FFEWS has four boundary conditions at Pafuri and Zongoene for upstream and downstream points on the main Limpopo River and the other two are Massingir Dam on Olifante River and Majacaze on Changane River. The models’ boundaries at Pafuri, Massingir and Maqueze meant that the actual flow observed and reported, as opposed to the forecasted flows, formed the values of the system at these points and at Majacaze. The system was designed like this because there were no data supplies upstream of Pafuri, Massingir and Majacaze to be able to forecast flows at these points.

25

Figure 2-1 Limpopo Telemetry Network Implemented by ARA Sul as Planned

Source: Personal Communication with Officials of ARA Sul It should be noted that:

• missing data at any one of these points means that the MIKE 11 FFEWS cannot be used;

• The accuracy of MIKE 11 flood forecasting was greatly compromised by the data and information shortfalls that include the long distance spacing of cross sections used in the flood routing model. These cross-sections’ locations and distances are described in Table 2.3, which indicate cross sections were widely spaced instead of intervals equal to or less than the active channel width during flood flows; and,

• MIKE 11 could not work for a long time and was subsequently abandoned as early as 2006 or thereabouts due to frequent shortage of data at the four boundary condition points and in a number of the telemetry stations.

26

Figure 2-2 Telemetry System for Limpopo River FFEWS as Constructed

Key 

  Base Station/Centre   Radio Relay Stn 

  Flow stn   Water level  

  Water level stn   Radio Repeater Stn 

  Weather/Rain Stn  

Table 2-3 Limpopo FFEWS Cross-section Location and Distances Apart

Upper End Point Upper End Point Distance Apart in Km

Basin Area above Upper point in km2

Comment

Pafuri Combomune 100 235, 930 Combomune Mabalane 100 257, 200 Mabalane Macarratane 70 259, 200 Macarratane Chokwe 15 - Near Chokwe Chokwe Sicacate 15 342, 000 Sicacate Chibuto 60 406, 210 Chibuto Xai-Xai 60 - Near Xai-xai Xai-Xai Zunguene 20 407, 970 ARA Sul took advantage of the satellite based rainfall estimates generated by SAFFGS and other weather forecasting centres and availability of USGS GeoSpatial Streamflow Forecasting Modelling, GeoSSFM, to forge a convenient flood forecasting system for the Limpopo River. With the availability of rainfall data in South African part of the catchment and Changane sub-basin, ARA Sul was able to calibrate the streamflow model with boundary conditions at Pafuri and Zongoene. Like the MIK 11 FFEWS, the GeoSSFM is also using the above cross sections in its flood routing. The GeoSSFM is also a convenient flood forecasting system for that part of Limpopo River in Mozambique and gives indicative flood forecasts with significant uncertainties due to its nature, use of satellite rainfall estimates that cannot be verified with ground truths use of overstretched reaches in routing and setting boundary conditions within the river itself with significant and huge areas outside its forecasting range.

Source: modified from personal Communication with Officials of

27

These setbacks in the existing FFEWS for Limpopo River need to be addressed in the establishment of an alternative system that better meets the stakeholder requirements and the hydrological characteristics of the Limpopo River. The reference point needs to be based on the opportunities that exist already and the ideal FFEWS available in the world today. This should include the WMO advances on sharing real time hydro-meteorological data and information using the WMO Integrated Global Observation System (WIGOS) and WMO Information System (WIS). The ideal or model FFEWS are discussed in section 2.2 in order to appreciate the recommendations for improving Limpopo FFEWS in chapter 3.

2.2 MODEL FLOOD FORECASTING AND EARLY WARNING SYSTEMS It is now realised that ideal river flow and flood forecasting and early warning systems should be those that have effective and efficient regional (basin wide) hydro-meteorological monitoring, scientific data analysis and forecasting models at an appropriate centre producing timely warning and forecast products, as described by Figure 2.1. The systems comprise reliable and rich data and information sources; the forecasting centre and flood areas linked with real time communication to enable operations for flood forecasting models save lives, protect property and infrastructure from destructions of floods. This is further discussed in the following sections on Real Time Flood and Related Data Monitoring; Flood Forecasting Models; and, Flood Early Warning and Response, with Practises in Trans-boundary FFEWS Management being the last one.

2.2.1 Real Time Flood and Related Data Monitoring

The real time data requirements for an ideal FFEWS is huge due to the need of having models that clearly represent what is happening on the ground, mathematically. The richer the data and information the finer are the river flow and flood forecasts. This is why river basins with rich real time data and information have accurate FFEWS.

The establishment of rich databases requires complex real time hydro-meteorological monitoring and data acquisition techniques and technologies – telemetry, radar and satellite data acquisition systems. The monitoring is needed before, during and after forecasting to gather historical, operational and verification or updating data and information, respectively. The historical data is needed for calibration and testing of the models while operational data is for carrying out the actual forecasting. The verification data is used to check and update forecasts. The data itself is collected and gathered from river gauging stations; weather and rain gauge network; and, radar and satellite measurements of rainfall and other meteorological data. The collected data is transferred to the regional forecasting centre via satellite, telephone (GPRS, CSD, etc.) and/or radio in real or near real time. The gathered data and information is processed and put in form and format required, such as SHEF, ready for forecasting or as part of forecasting. This is done at the forecasting centre itself as it hosts PCs that process, analyse and manage data and run flood forecasting models, accordingly. There will always need to be a balance between the amount of data collected from both a cost and effectiveness perspective and the quality and accuracy of the resulting flood forecasts.

28

Figure 2-3 Model Flood Forecasting and Early Warning System

Source: Modified from personal communication with USGS Personnel

2.2.2 Flood Forecasting Models

A number of efficient and effective flood forecasting models have been developed and put in place in river basins that are flood prone across the globe, with an aim of saving lives and protecting property and general infrastructure vulnerable to flood disasters. However, very few have withstood the tests with technical and administrative support from their vendors. The NWRFS and the MIKE 11 Flood Watch or its upgrades are among those that have withstood test of times and are well supported by their vendors. The NWRFS is a robust state of the art river flow and flood forecasting system complete with hydrological forecasting and routing models and routines. The system includes data handling and presentation features of hydrological process that lead to forecasting hydrological event or events. As mentioned in 2.1.4, NWSFRS require state of the art PCs to handle its large data storage requirements and perform complex calculations in its components’ modules. It has reached this level of complexity and performance after over 20 years of its use and is being constantly refined and improved. The model’s wide testing and use has not only been in USA but also in other countries throughout the world including Peoples Republic of China, Panama, Republic of South Africa, Nicaragua and El Salvador The NWSRFS is a modular build model and structured to execute interrelated software procedures that perform a wide variety of hydrological/hydraulic and data management operations. Operations in NWSRFS are a set of equations of motion governing the flow of water through a portion of the hydrological cycle. There are also operations to display results, or to perform utility functions. Typical operations are sub-models such as:

• rainfall/runoff models that include functions for soil moisture accounting, antecedent precipitation index calculations, etc.;

29

• routing models with functions including those for muskingum, dynamic wave routing model, etc; or,

• temporal distribution model with functions for calculating unit hydrograph. Figure 2-4 Three Operational Components of NWSRFS

The scientific algorithms for programming these operations are based on modular functions so that the functions can be shared unchanged, among major components of the NWSRFS. This also allows development of model components by a number of individuals and then combined into a total system at a later stage. The system is currently supported in Hewlett-

Packard UNIX and LINUX operating systems, with a MS Windows™

version also available but not supported by the NWS. The NWSRFS itself has three operational components – Calibration System, Operational System and Ensemble Streamflow Prediction (ESP) System – as shown in Figure 2.2.

The Calibration System generates time series based on historical data and determines the model parameters. The Operational Forecast System produces short-term river and flood forecasts using calibrated model parameters and maintains model state variables, while the Ensemble Streamflow Prediction System makes probabilistic forecasts extending weeks or months into the future using current model states, calibrated model parameters, and historical time series.

30

On the other hand, MIKE FLOOD WATCH is also a robust decision support system for flood forecasting and early warning and real time operations of water resources infrastructure systems. This and its GIS capabilities have made MIKE 11 and MIKE FLOOD WATCH, FW, gain significant recognition and widespread use throughout the world since some 20 years or so ago. The FW and its open GIS based environment capabilities enhance integrating real time data management, data analysis, data presentations, flood forecast modelling and dissemination methodologies into a flexible powerful tools with features that include:

• user-friendly client server solution-comprehensive and highly visual presentation of planning, real time and forecast information within the GIS platform and on the web;

• Data import capabilities – complete range of planning and real time data and information, including maps and mapping, satellite/radar imagery, telemetry data, meteorological forecasts, etc;

• Advanced data management that makes it able to have complex data storage (geo-database), complicated data analysis (quality assurance/quality control handling, interpolation, extrapolation, gap filling, rating curves, time series statistics), data presentation, etc.;

• Stable and resilient system, including a data hierarchy that ensures robustness to poor or missing data;

• Model and database management shown in Figure 2.3, that enables it accommodate and execute different model types from different model suppliers including DHI, HEC, NWS and Sacramento models, etc.;

• Flexible scenario management that enables it to, for example, optimise strategies for reservoir release operations or comparative evaluation and assessment of expected areas under flood inundation, levee failures, etc., which are further used in establishing target areas for flood response and related remedial measures and issue warnings, accordingly; and,

• Versatility in dissemination management that is used in issuing warnings and distributing flood forecasting products through e-mails, fax, web/internet, emergency telephone number calls and telemetry databases as well as SMS on cellular phones and other devices that receive and send SMS messages.

The MIKE FLOOD Watch system require FW central database server, front end and back end array of PCs. The central database server comprise database server (e.g. INFORMIX, ORACLE, SQL SERVER) or file based database, which gets input from an array of front end PCs (telemetry database and a cluster of model servers). Each front end PC constitutes the main entry point for forecasting personnel and operates under Windows XP Professional OS. On the other hand servers at the back end of the system run Windows 2003 server and these comprise FW server and a cluster of servers for task executions (e.g. simulations). All definitions and tasks programmed by individuals from front end PCs are stored in FW central database, which means that all the information, such as real time data, scenario definitions, simulation results and other forecasted results and products are accessible from front end PCs regardless of which PC was used in deriving the data and information. There is, therefore, full sharing of real time data, models and simulations by the forecasters.

31

Figure 2-5 MIKE FLOOD WATCH Infrastructure Framework

The FW system is developed to interface, manage and operate different model codes and types in a generic manner based on industry standards. This is made possible by its model code data structures and model engine executables availability and accessibility that provide the appropriate interface (data bridge) to FW. The FW is, therefore, ideal for establishing FFEWS where at least two base station or forecasting locations/centres are required, as it is the case with ARA-Sul MIKE II system. It is also appropriate for situations where details of forecasts, flood hazard maps and other FFEWS products are required and disseminated throughout the basin.

2.2.3 Flood Early Warning and Response

Flood early warning is the end of an FFEWS in ideal early warning systems and comprises flood watch, advisories and warnings. Flood watch is the vigilance and preparedness for the likelihood of flood event(s) occurring. This takes place when forecasters have information that shows conditions are ripe for river floods or flash floods.

32

Figure 2-6 MIKE FLOOD WATCH Interfacing with Model Tool

The vigilance also involves forecasters issuing instructions or advisories to the vulnerable communities and general public to monitor news and be prepared to act when told to evacuate, move valuables to higher ground, close bridges or roads, etc. The flood advisories can also include issuing statements such as calling the vulnerable communities and general public to monitor their surroundings, watch for warnings on radio, television or even SMS messages. Depending on circumstances, the vulnerable communities and the general public are advised:-

• To listen and continue listening to radio messages and further warnings and advisories;

• Where safe places are and asked to move to or how to identify safe place where to move to;

• What to stock pile and carry (food, medicine, water, clothing, toiletries, etc.); • To evacuate from potential and likely to be flooded areas; and, • To be prepared to leave when told to; etc.

Flood warnings themselves are issued when a flood event has been predicated and should comprise statements on where and when the flood will occur, its impact area, magnitude and duration. These warnings are issued to Disaster Management authorities, vulnerable communities and general public. The advisory statements also include advisory or instruction statements crafted along the phrases that include:

• Urging the vulnerable communities and the general public to be alert and vigilant during the flooding;

• Advising on things to do and not to do (such as do not drive unless you have to, drive with enough fuel, do not drive in rising waters over the roads, when caught in flooded area abandon vehicles (if water surround the vehicle or vehicle stalls) and urgently climb higher grounds, buildings, etc., do not drive flooded roads, avoid loitering in disaster areas to reduce dangers to human beings or give space to emergency operators, etc) among the vulnerable and general public in the flood area;

33

• Requesting vulnerable communities and general public to be on the watch out for washed away roads, bridges, power lines; etc., and report to the authorities (giving telephone numbers to contact);

• Reverberate and urge reporting response to flood warning, (which should include periodic repetition of flood warnings, indicating flooding has already occurred or updating information on new dangers, revisions of flooded areas, rescue efforts done or planned, situation in temporary relief camps, schools, health services, etc. during the flood events and other information,) after the first warning was issued;

• NEVER drive through flooded roadways but stop and turn around to avoid drowning because: -. o The roadbed may be washed out and can cause pot holes or gullies that

can invisible from water surface but may stop the vehicle or make it tip over;

o Control loss of the vehicle in only a few inches of water that can make driving in water fatal;

o Vehicles can be swept away by less than 0.6 metres of water as the cars easily loss buoyancy; and,

o A barricade has a purpose and should not be disregarded and, therefore, obey instructions, turn around and go through another way!

• General Warning should advise vulnerable communities and general public to get to high ground and climb to safety with phrases that urge vulnerable communities to o Get out of low areas that may be subject to flooding; o Avoid already-flooded areas and not attempt to cross flowing water; and, o Stay away from power lines and electrical wires because of electrocution

threats. • Evacuation should be immediate whenever risk has been identified or

vulnerable communities and general public have been advised to that effect and should do so because it is necessary to: o Act quickly (urging individuals to save themselves first and their

belongings later); o Move to a safe area before access is cut off by rising water; o Families should use only one vehicle to avoid getting separated and

reduce traffic jams in built up affluent areas; o Shut off water, gas, and electrical services before leaving; o Secure homes and lock all doors and windows; and o If there is need for directing vulnerable population to a specific location,

there should be counter urging all the people there’ • Never try to walk or swim through flowing water or let the children wonder in

flood waters:- o Is flowing above ankles deep but instead STOP! Turn around and go

through another way; and, o Move swiftly since water 15 centimetres deep can knock a person (child)

off his/her feet; • It is important to communicate that raging flood waters are dangerous and

vulnerable communities and general public should:- o Be aware that people have been swept away wading through flood waters; o NEVER allow children to play around high flowing water, storm drains,

creeks, or rivers;

34

o Not go after the victim! But if possible throw the victim something to help them float, such as a spare tyre, large ball, or foam ice chest;

o Use a floatation device; and, o Call emergency telephone numbers for assistance and give the correct

location information for help to reach them. The operations for response activities such as evacuation, temporal resettlements, provision of relief items etc. that are done prior, during and after floods are not necessary the responsibilities of the forecasters (NHS or NMS) but disaster management authorities. However, forecasters’ professional knowledge of the floods and flooding characteristics impels them to continue informing the flood victims about what they can do and they cannot do after the floods, such as: -

• Informing victims and general public to wait until it is safe to return and urging them to continue to: o Monitor local television and radio stations for any updates; o Not to return to flooded areas until authorities indicate it is safe to do so;

and, o Not visiting disaster areas following a flood, as their presence may hamper

urgent emergency response and rescue operations; • Continue maintaining vigilance while travelling in the flooded areas by:

o Following recommended routes and NOT to sightsee; o Watching for washed out roads, landslides, downed trees or power lines

and reporting them to authorities accordingly; and, o Staying away from downed power lines.

2.2.4 Trans-boundary FFEWS Management Practices

The practises discussed from 2.2.1 to 2.2.3 are normally done by responsible national agencies when the rivers are not trans-boundary, unlike the Limpopo River which is a shared watercourse. In shared watercourses, the activities 2.2.1 to 2.2.3 are effectively and efficiently performed by a basin agency cooperating with the NHSs and NMSs of all riparian countries. The management, operation and maintenance of the telemetry network and flood forecasting system is coordinated by basin agency while national agencies (NHSs and NMSs) tend to receive and issue flood warnings. This is the practise in most trans-boundary river basins, with either a regional flood forecasting centre or coordinated river flow forecasting.

There are number of shared watercourses where FFEWS are being implemented but the regional practices in Mekong and Rhine Rivers provide the best examples. FFEWS in Rhine is basically done through International Commission for the Hydrology of the Rhine, CHR, which is an organization in which the scientific institutes of the Rhine riparian states develop joint hydrological measures for sustainable development of the Rhine River Basin. Under the Commission scientific institutes such as hydrological services of member states freely share data and information for the entire Rhine River Basin. This allows the hydrological services institutes and agencies interested or responsible for flood forecasting and early warning to forecast Rhine River flows, including floods and disseminate flood forecasting and warning products as discussed in 2.2.2 and 2.2.3. The abundance and availability of river flow and other hydro-meteorological data and information and communication systems in all riparian states makes this sharing in real time easy and common practice.

35

In the Mekong River Basin, the development of Mekong River Commission (MRC) and its flood forecasting and early warning system have been centred on the following:

• In 1995 the MRC was created under an agreement among Cambodia, Laos, Thailand, and Vietnam to promote “… the Sustainable Development of the Mekong River Basin”.

• In 2000 o 50-year floods killed 800 (mainly children) and caused $400 million (US) in

damages; and, o The MRC Council—ministerial-level policy making body—directed the

MRC Secretariat to prepare a Flood Management and Mitigation (FMM) Strategy.

• In 2001 the Council adopted a FMM Strategy which centred on implementing flood preparedness, mitigation, response and recovery. This included the development of the Mekong River flood forecasting and early warning system via a highly participatory process, targeting and emphasizing on the following: (1) Providing technical products and services from FFEWS such as those

discussed in 2.2.2 and 2.2.3. (2) Addressing differences and facilitation among the member states in

development, operation and maintenance of telemetry systems and flood forecasting models; and,

(3) Capacity building and technology transfer among the countries and their participating agencies dealing with flood preparedness, mitigation, response and recovery.

The flood preparedness, mitigation, response and recovery strategies called for measures that include land use planning measures, structural measures, flood preparedness measures; and, flood emergency measures with priorities set as presented in Table 2.4.The Strategy was approved and is being implemented under the MRC Flood Management and Mitigation Programme, FMMP, which has five components as follows:-

1. Establishment of a Regional Flood Centre where staff from the member states are seconded to and participate in operation and maintenance of the Mekong River flow and flood forecasting and early warning system using MIKE II and it upgrades

2. Structural Measures and Flood Proofing such as dams, dykes, levees, river training or flood proofing with house on styths or stands

3. Mediation of Trans-boundary Flood Issues 4. Flood Emergency Management Strengthening with necessary capacity building

among member states; and, 5. Land Management to promote soil and water conservation and reduce excessive

run-off during storms Since adoption of the FMM Strategy the MRC (with donor support) has:

• Established a FMMP Centre in Phnom Penh, Cambodia; • Conducted capacity development and technology transfer programmes; • Improved data-collection and flood-forecasting technology; • Installed a Flash-Flood Guidance System; • Conducted Annual Mekong Flood Forums where flood issues are discussed,

shared and lessons learnt and consolidated;

36

• Established facilities, created access data bases, and adopted decision-support tools for sustainable development in each country; and,

• Coordinated the provision of flood information such as that discussed in 2.2.2 and 2.2.3 to vulnerable communities and general public.

Table 2-4 Table of MRC Flood Risk Management Strategies and Priorities

Role Element Providing Technical Products and Services

Addressing Differences and Facilitation

Capacity Building

Land Use Planning Measures

Structural Measures

Flood preparedness Measures

Flood Emergency Measures

KEY HIGH PRIORITY MEDIUM PRIORITY LOW PRIORITY

The Mekong FFEWS was developed and is being implemented in a policy and legal environment similar to that offered by LIMCOM and can be established in the Limpopo River basin. The Rhine FFEWS model would have problems to operate in the basin as the real time data and information available is inadequate and standards uniformity among the riparian countries is absent too. It is important, therefore, that improvements to Limpopo River Basin FFEWS consider the existing opportunities inside as well as outside and exploit the cooperation among riparian states accordingly, like what the Mekong Commission Council did.

37

Chapter 3 : RECOMMENDATIONS FOR IMPROVED LIMPOPO FFEWS The recommendations for improving flood forecasting and early warning system for Limpopo River should be based on the experiences of that developed and used by the Mekong River Commission. In this regard, the recommendations are centred on improvement of monitoring systems; consolidation of FFEWS institutional development and capacity building. Cost estimates for improving the Limpopo flood early warning system are also provided. These recommendations should be the basis of designing, constructing and installing the Improved Flow/Flood Forecasting and Early Warning System for the Limpopo River Basin. Some detailed survey and inspection of the telemetry stations proposed and assessment and identification of detailed needs would also be required to update this proposal. The updating of the designs should include the detailed cost and financing of implementing the system including capacity building. This would essentially be updating section 3.1 to 3.4 below.

3.1 IMPROVEMENT OF LIMCOM MONITORING SYSTEMS

3.1.1 Consolidation and Modernizing Telemetry Network The selection of SADC-HYCOS stations in Limpopo River Basin in South Africa, Botswana and Zimbabwe appear representative of the catchments in the respective countries. These stations would complement the automatic weather stations within and in the vicinity of the respective catchment portions in gathering real time data and information for river flow and flood forecasting and early warning system, if properly improved, operated and maintained. The SADC-HYCOS stations in Mozambique are only two and, therefore, not representative. The network used by ARA Sul Limpopo River FFEWS, however, is representative enough and it is recommended that this should be part of the consolidated and improved LIMCOM FFEWS telemetry component. It is, therefore, recommended that the network listed in Tables 2.1 and 2.2 be the consolidated telemetry network for the Limpopo River FFEWS. It is further recommended that the consolidated network should be modernised and made robust with rehabilitation/upgraded design, construction/rehabilitation, operation and maintenance that incorporates:

(i) Upgrading and installing satellite based data transmission systems at each station. The SADC-HYCOS stations need to be installed with Phase II equipment or better and the ARA Sul stations should be upgraded to have satellite transmission while keeping the radio system. The radio transmission system should be rehabilitated with masts replaced by aluminium or corrosion resistant metal poles. Other data and information communication systems like GSM, GPRS and telephone lines should be promoted and introduced at each station, wherever possible and particularly key stations, to increase to increase chances of collecting and sending data to base station, at all the times;

(ii) Replacing solar powered batteries with long lasting stand alone batteries with life span of at least 10 years; such as those provided by Solinist Canada Limited (http://www.solinst.com/Text/text-dataloggers.html), or Coronics (www.coronis.com/en/faq.html) or Dynamic Logic

38

www.dynamiclogic.co.uk/water/qg_out.html). Investigations are needed to make sure all the stations can have such batteries with the existing equipment. The equipment that fails to accommodate the long lasting batteries should be replaced with equipment which can accommodate the long lasting batteries;

(iii) Equipment and accessories used should all have local technical support services in supplying spare parts and servicing the components, in each participating country; and,

(iv) The stations themselves should be made vandalism proofed and maintenance guaranteed with batteries and instruments well protected. The rehabilitation of the stations should take advantage of long lasting batteries and watertight/corrosion free casings to provide equipment housing that are secure and protected from vandalism, flood damages, corrosion and theft. The cause stations should be upstream of bridges to maintain and operate stable rating curves for long periods and reduce risks of discharge stations being redundant during flood period due to possible changes of channel configuration.

3.1.2 Harmonizing Data Collection and Processing Routines

The consolidated telemetry network should be supported by a strong and harmonised data collection and processing procedures, guidelines and standards. This should also involve modernizing data collection including river discharge measurement that employ modern equipment and methodology like using acoustic Doppler systems for measuring water velocities with hand held, winch held or boat mounted devices. A manual of standards and guidelines on data collection and procedures, standards and guidelines for all required activities should be developed (based on existing materials) and used and should cover areas, including:-

• discharge measurements; • time and intervals of recording rainfall, water levels, evaporation, etc.; • computing discharge measurements and averages of discharges/flows; • installation and maintenance of gauging stations, rain gauges, etc.; and, • data reporting and archiving formats, etc.

Guidelines and standards on harmonization of data collection and processing procedures will also include frequency and details (methods used including slope area surveys) of discharge measurements. The routines for station inspections, data entry, database maintenance, etc. should also be included in the manual. This manual should essentially be highlighting and tailoring all the required activities for data collection, processing, analysis, archiving and dissemination outlined in the WMO guidelines and standards. It should be part of the overall manual for improved Limpopo River Basin FFEWS.

3.1.3 Development and Launching Data Sharing Systems

The four countries require sharing data and information for the consolidated and improved Limpopo FFEWS development and operation. The existing data sharing arrangements between Mozambique and South Africa should be extended to and among all the riparian states. The sharing of hydro-meteorological data and information should be within the

39

country between NHS and NWS as well as among the riparian states between the NHSs and the regional river flow/flood forecasting centre. Appropriate data and information communication hardware and software should be procured, installed and commissioned with corresponding protocols for transmitting and disseminating data and information between NHSs and NWS and countries and the FFEWS centre. GTS, WIGOS, WIS and their accessories hardware and software available should be assessed, selected and installed to enable sharing data and information with the FFEWS centre and countries. The basin wide data sharing system should be tested, commissioned and used during and after the project.

3.1.4 Trans-boundary Telemetry Network Management

The Recommended Limpopo River Basin FFEWS telemetry network of stations is in the jurisdiction of four riparian states with different means of operation and maintenance. However, to ensure an efficient and effective system, these Limpopo River Basin stations should be operated and maintained in a consistent and commonly agreed manner. To avoid disparities in operation and maintenance, LIMCOM should oversee and audit the maintenance programs to ensure the same level of operation and maintenance services in each country. During the project period, an assessment should be made to determine levels of LIMCOM support for operation and maintenance of the telemetry system and how this will be implemented, including for the base station and sub-base stations in each riparian country. The level of support should also be extended to the hardware and software installed at the base and sub-base stations including the hardware and software for the consolidated Limpopo River low/flood forecasting and early warning system.

3.2 CONSOLIDATION OF RIVER FLOW FORECASTING AND EARLY WARNING

3.2.1 Objectives and Outputs of the River Flow Forecasting System

The problems of floods and drought disasters, high competition for water and threats of water scarcity, water pollution and water utilisation conflicts in Limpopo River, as discussed in section 1.3 entail that the objectives of the Limpopo River Flow/Flood Forecasting and Early Warning System should be to provide decision support in:

• Issuing flood warnings intended to save lives, property and public infrastructure from flood losses and damages;

• Conserving water resources and protecting its infrastructures from flood damages by efficiently regulating flood releases from such infrastructures; and,

• Advising authorities on equitable and reasonable water allocation and water pollution control with the knowledge of real time water availability in rivers and streams of Limpopo watercourse.

This decision support should essentially be a River Forecasting and Early Warning System, RFEWS for the Limpopo River Basin. Such a system should fulfil these objectives and respond to LIMCOM’s aspirations on disaster, water quality and water allocation management as discussed in second paragraph under Background chapter.

40

The expected outputs are the

• Regional RFEWS Centre and sub-centres in other countries with legal personality, fully equipped and functioning;

• Regional Telemetry system rehabilitated, installed, developed, commissioned and functioning;

• Real Time Limpopo River Basin RFEWS commissioned and functioning; and, • Trained Personnel at the RFEWS Centre and sub-centres in each riparian

country.

3.2.2 Establishment of Regional River Flow Forecasting Centre

The RFEWS Centre should it be agreed too, would be hosted within one of the existing institutions in the LIMCOM member states. The host institution, that is willing, has characteristics suitable for accommodating RFEWS centre within the basin, needs to be selected, adopted and secured using selection criteria and negotiated agreements between the institution, LIMCOM and the riparian states, to ensure ownership of the Centre by the riparian states.

The selection would start with a survey and assessment of the potential host institutions in the four riparian states. Criteria would be developed to guide assessment and selection of the host institution, which will be recommended to the LIMCOM, accordingly. LIMCOM will then approve the host institution.

Once the selection has been completed, special agreements will be prepared between the host institution and its Government on one hand, LIMCOM and the riparian states, setting out roles, responsibilities and overall governance. The riparian states will have an agreement on how to fund and run the RFEWS centre plus the LIMCOM telemetry system, including general operation and maintenance of the system and its joint data collection programmes. All agreements will have to be drawn during early stages of the project in order to facilitate project implementation and will have to be signed at the highest possible level to support sustainability of the project.

3.2.3 Development/Commissioning River Flow Forecasting Models

It is recommended that the design and improvement of the River Flow/Flood Forecasting and Early Warning System for the Limpopo River Basin should be based on MIKE 11 (FLOOD Watch) installed at ARA Sul in Mozambique. During the design of the RFEWS, serious consideration should be given to upgrading the MIKE 11 software to higher and more recent versions of MIKE FLOOD WATCH. The upgraded version should be complemented by other flood forecasting models like the Geospatial streamflow models and flash flood models in sub-catchment areas where MIKE FLOOD WATCH cannot work due to inadequate data and information.` To accommodate and improve forecasting results required to give flood warning and decision support to LIMCOM, there is need to:

41

• Have cross-sections at appropriate intervals throughout the main river, particularly in the flood prone areas of Lower Limpopo River in order to generate adequate forecast details on the extent and depth of flood waters;

• Design the RFEWS with alert levels targeted at flood watch and warning levels of various thresholds depending on the need of vigilance and area, water resources, roads, railway and other public infrastructures, population, settlements and property to be affected;

• Design the river flow forecasting component with alert levels targeted at high and low flow levels of various thresholds depending on the need of vigilance needs, water abundance or scarcity, pollution and user conflicts;

• Link the flood warnings and flood warning thresholds to various flood zone maps and the developments to be affected;

• Establish flood zone maps at various water levels along the flood plains associated with flood warning thresholds for an area or areas and water resources, roads, railway and other public infrastructures, population, settlements and property to be affected; and,

• Develop remedial measures and response mechanisms to safeguard lives and protect water resources, roads, railway and other public infrastructures, population, settlements and property to be affected.

The needs for low flow and water quality river forecasting should equally be addressed in the design, development, operation and maintenance of the RFEWS. There will be need to critically examine these needs and incorporate them, in a balanced manner, accordingly.

3.2.4 Manual of Standards and Guidelines of Limpopo RFEWS There is need for common standards and guidelines to be practised in running the RFEWS to assist in coordination and synchronization of various activities and ensure accurate and similar results throughout the basin. This is because of the Limpopo River RFEWS is expected to be operated and run in each of the four riparian states and, should it be established, at the RFEWS Centre, with complex demands and complicated data collection, processing, analysis, archiving and retrieval. Data quality control and integrity of flood warnings from flood forecasts generated by several people at various places can only be upheld when a common manual of standards and guidelines of all practices are available and used by everyone involved.

The major activity under this output is preparation of a consolidated Limpopo RFEWS Manual of standards and guideline practices for its telemetry, flood forecasting models and flood warning procedures (based on available guidance and standards). The telemetry standards and guidelines need to cover operation and maintenance of the equipment and data capture, storage and communication software. The standards and guidelines for operation, maintenance and updating of flood forecasting models should be documented and those for flood forecasting, flood watch vigilance, flood warning and advisory statement crafting need to be included for early flood warning procedures and activities.

42

3.3 RFEWS INSTITUTIONAL DEVELOPMENT AND CAPACITY BUILDING

3.3.1 Technical Capacity Building Programme Development

There is need for assessment, evaluation and identification of capacity building needs at RFEWS centre and national RFEWS centres and preparation of capacity building and training programme. This will target identification of policy and institutional reforms, technical facilities and human resources needed at each centre or riparian state hat would be required for smooth operation, maintenance and updating the RFEWS.

The policy and institutional reforms will include changes needed in financing arrangements, administrative and institutional developments needed for operation, maintenance and updating of the various components of the RFEWS. The instruments for implementing such changes will also need to be developed and recommended to LIMCOM. The technical facilities will include tools, assets and implements that would empower the centres to develop, operate, maintain and update the RFEWS. The recommendations should include financial implications of the listed tools, assets and implements. The LIMCOM will use these to empower the centres, accordingly. The human resources assessment and needs identification will include training needs. The needs identification will culminate in preparation of training programme for training staff of the centres in operation, maintenance and updating the RFEWS. These will be recommended to LIMCOM for approval and adaptation.

3.3.2 Capacity Building and Training

The full resourcing and empowering of the regional RFEWS and its sub-centres in each country is strongly recommended. This will be achieved by implementing the recommendations in the institutional development and capacity building programme proposed in 3.3.1. It will involve procurement, supply and equipping the RFEWS Centre and National centres plus conducting in-service and professional staff training for both regional and national RFEWS Centres’ staff. The training programme implementation will aim at developing skills among staff so that riparian staff members participate in the design, installation, calibration and commissioning of the telemetry systems as well as RFEWS. The development of the training materials will be undertaken during the early part of the project and build on available materials. The materials would be developed and presented in way which makes them easy to incorporate into the manuals to be developed under section 3.1 and 3.2. The latter training materials will be based on the manuals for standards and guidelines themselves supplemented by other guidance material as necessary.

3.4 WORK PROGRAMME, INPUTS AND COST ESTIMATES

3.4.1 Work Programme and Input Requirements

It is expected that the project period will be two and half years with the main activities lasting for 18 months. This assumes that the: -

43

(i) Civil works and recorder houses already completed or will be constructed by the riparian countries and that trained personnel from riparian states will play a critical role of providing support in the installation operation and maintenance of the telemetry equipment in the field and at national RFEWS sub-centres;

(ii) Offices for the Regional River Forecasting Centre (should it be established) and its sub-centres in the riparian countries and regional and national or local counterpart staff will be provided for by LIMCOM and riparian countries, respectively; and,

(iii) The development of the flood forecasting and early warning system will be contracted to a consultant/contractor and will implemented over a two year period to allow for calibration of the models; and,

(iv) The main activities of procuring and/or rehabilitating, installing and commissioning of the telemetry system, establishment of RFEWS centre and sub-centres and procurement and initial supervision of the RFEWS consultant/contractor will be done by the Project Manager and Project Engineer, who will be regionally recruited.

The proposed arrangements also assume that most of the telemetry equipment installation works will be carried out by the trained staff in each country, under the supervision of the Project Engineer. In this regard the work programme activities include:

(i) Preparation of an Inception Report by the Project Team (ii) Procurement of telemetry equipment and establishment of the RFEWS Centre

and sub-centres; (iii) Rehabilitation, installation and commissioning telemetry system; (iv) Procurement of modelling consultant/contracts (v) Supply, installation and Commissioning the RFEWS by contractor; (vi) Preparation of the RFEWS Field Manual (vii) Capacity Building and Training.

The work plan in Table 3-2 is indicative and it is expected to be updated improved greatly in the Inception Report that the Project Team comprising Project Manager and Project Engineer will produce and submit within three months of the commencement of the project. Once the Inception Report is approved the procurement of remaining telemetry equipment and modelling consultant/contractor should commence. Table 3-1 Work Plan

Period of Implementation in Month from Commencement Time Item Activity 3 6 9 12 15 18 21 24 27 30 i Inception Report Preparation ii Procurement of Telemetry Equipment iii Establishment of RFEWS centres iv Procurement of Modelling contractor v Rehabil. Installation of telemetry Stations vi Preparation of Telemetry System Manual vii Installation and Calibration of RFEWS viii Preparation of RFEWS Manual ix Capacity Building and Training x Commissioning and Handover of RFEWS The Project Manager and Engineer are the envisaged technical assistance provided under the project, which will comprise the core Project Team. As part of Inception Phase of the

44

project, the two would assess and update this proposal particularly determining the exact number of telemetry stations requiring complete new equipment, which ones need repairs and upgrading and listing the requirements accordingly. The team will also assess the installation requirements for the stations that need new equipment or rehabilitation or upgrading and the equipment housing requirements in order to revise the cost estimates accordingly, as part of Inception Phase. The Team will also assess the technical capacity of the National Hydrological Services in the installation and upgrading of the telemetry stations and update input requirements, accordingly. It will carry out preliminary assessment of the existing institutions and prepare criteria and guidelines for selecting the Regional River Forecasting Centre during the Inception Phase. The Inception Report would be submitted to LIMCOM and WMO for approval, accordingly, within three months of the project commencement.

The work would proceed with the procurement of the telemetry equipment; establishment of RFEWS Centre and its sub-centres; preparation of the telemetry manual and capacity building of the centres and training riparian staff in installation and operation of the telemetry equipment in the second quarter of the project. This will allow countries to participate in rehabilitation and installation of the telemetry stations in the third quarter of the project. It will also allow procurement of the modelling consultant/contractor in the third quarter of the project, which will in turn allow supply, installation and calibration of the RFEWS models in the fourth quarter in the first year to third quarter after the second year of the project. The contract of the modelling consultant will include supply of the models as well as licensing and all the activities in 3.2 plus those in 3.3 related to RFEWS modelling, early warning and capacity building and training. The work plan for capacity building and training would, therefore, be adjusted as soon as the contract with the modelling consultant has been known. It is envisaged that the work of coordinating and managing the project would be significantly reduced after models’ installation and initial calibration work which are expected to be completed in the sixth quarter or eighteen month from the beginning of the project and regional counterpart staff would takeover. The services of the Project Manager and Project Engineer will be terminated at this point. The contract for the modelling consultant will continue with models’ calibration, manual preparation, training regional and local counterpart staff and commissioning and handing over the RFEWS.

3.4.2 Budget and Cost Estimate of the RFEWS The basis of the budget is the provisional list of stations requiring new equipment, which has been put at 25, comprising 9 stations which currently have no equipment in Mozambique, 10 estimated to be needed in South Africa and half of those targeted in Botswana and Zimbabwe for the purpose of estimating telemetry cost including spare parts for rehabilitation and upgrading. This is in addition to the SADC-HYCOS Phase II equipment not yet installed. Besides, each RFEWS centre will need one high powered PC and 4 work stations and accessories and flood forecasting model licences. The budget is also based on the capacity building and training that is expected to require three regional workshops of training and sharing experiences. It is estimated that each country will need 10 participants drawn from National Hydrological, Meteorological, Disaster

45

Management, Civil Defence Protection and universities. The training workshops are expected to last for 3 days each. There is also need for two double/cub vehicles for project transport for visiting countries for the consultations and stations inspections during Inception Phase and also during rehabilitation and installation of the equipment.

Considering the input requirements listed above, the preliminary cost estimates for improving the River Flow and Early Warning System for the Limpopo River Basin is estimated at United States Dollars three million (US$ 3.0 M). The budget and cost estimate of the input has been itemized and estimated accordingly as per above inputs, as follows:-

(i) 25 no. of telemetry equipment sets @ $ 4,420 each; 110, 500 (ii) Discharge measuring set (equipment & accessories (4 sets)) 316, 800 (iii) RFEWS development contract

a. 16 no PC work stations and accessories 36, 000 b. 4 no. PC servers – one for each RFEWS Centre 16, 000 c. Modelling Software and licenses purchase (4 sets) 580,000 d. 36 man-month consulting services @ $22,000 792, 000

(iv) Project Manager & Project Engineer 486, 000 (@ $15,000 & $12,000/mm for 18 months) (v) Project Vehicles – two d/c pick-ups @ $61,000 each 122, 000 (vi) Vehicle running costs at $9,500 per month field work 171, 000 (vii) Office Running Cost -$ 8,500 per month for 18 months 153, 000 (viii) Workshops and Training transport and accommodation, etc. 164,600

Note: the above costings are based on the estimates per station as follows:

Water Level Datalogger with 8 to 10 year battery 1,670.00 Water level data logger accessories 530.00 Rain Gauge datalogger and accessories 595 Air Temperature/Humidity, Sensor with replacable head 480 Water Quality probe set 740 Subtotal 4,015.00 10% contingency 401.50

Total 4,420.00 Also note that the discharge measurement set and cost estimate comprise:

Boat discharge measuring equipment (based on Sontek River Surveyor)

35,200.00 Handheld or cable car fitted discharge measuring equipment (based on Sontek flow Tracker)

13,400.00

Dingy Boat 4 m by 1.95 m with 25Hp engine and trailer for discharge measurement

18,450.00 Spare 25 hp outboard engines 6,600.00 Spares 2 m wading cable complete reel switch & jack-plug connector. 250.00 35 m suspension/signal cable fitted with Model 001 reed switch & jack – plug connector 1,550.00

  contingency 5%

3,772.50

      SUBTOTAL

79,200.00

46

The capacity building and training budget workshop has been estimated as follows 45 participants at average of $450 an airticket for three times 60, 750 45 participants for 4 days at $150 ad day DSA for three times 81, 000 Conference charges, stationary and handout $5,000 a workshop 15. 000 5% contingency 7, 838 SUBTOTAL 164, 600                

47

REFERENCES 1. LIMCOM Limpopo River Awareness Kit by Limpopo Watercourse

Commission Secretariat, Maputo, http://www.limpoporak.org/ 2. DNA National Water Resources Management Strategy for

Mozambique, DNA, Maputo, Mozambique 2005 3. World Bank Memorandum – The Role of Water in Mozambique Economy

–Identifying Vulnerability and Constraints to Growth, World Bank, Washington DC, USA, 2005;

4. World Bank Country Water Resources Assistance Strategy for Mozambique, World Bank, Washington DC, USA, 2006;

5. INGC et al Atlas for Disaster Preparedness and Response in the Zambezi Basin, - FEWS NET MIND, by INGC, FEWS-NET & INAM, Famine Early Warning System Project Office, Maputo, 2011;

6. INGC et al Atlas for Disaster Preparedness and Response in the Limpopo Basin, - FEWS NET MIND, by INGC, FEWS-NET & INAM, Famine Early Warning System Project Office, Maputo, 2003

7. DNA Water Resources of Mozambique Synopsis of 1999 by DNA, 1999

8. UNDP Human Development Report 2006 – Beyond Scarcity: Power, Poverty and Global Water Crisis by UNDP, New York, 2006

9. Solinst Groundwater Monitoring, Water Level Meters, Dataloggers, Bladder Pumps, Peristaltic Pumps, Water Quality, Piezometers, Groundwater Samplers, Water Level Meter, Levelogger, Waterloo Hydro-geologic Software http://www.solinst.com/Text/texthome.html

10. SonTek River Surveyor/Flow Tracker - Discharge Measuring Current Meters, http://www.sontek.com/adp-acdp/

11. DHI MIKE FLOOD WATCH – Managing Real Time Forecasting by Claus Skotner, et al, DHI Water & Environment, Denmark, 2005 http://www.dhigroup.com/upload/publications/mike11/Skotner_MIKE_FLOOD_watch.pdf

12. DNA Joint Limpopo River Basin Study, Scoping Phase, DNA, Maputo, 2010.

13. NWS Susquehanna Flood Forecasting and Warning System, National Weather Services, USA, http://www.susquehannafloodforecasting.org/

14. E. Plate et al Early Warning System for the Mekong River by Enrich Plate & T. Insisiengmay, Science Press, New York, 2002