GNE: G lobal N EWS Modeling E nvironment What it is / what it’s not How it works Components, samples Installation NEWS 2 Implementation Emilio Mayorga May 2009 [email protected]http://www.marine.rutgers.edu/globalnews/GNE GNE Training Session 1 Content last edited August 28, 20
23
Embed
GNE: Global NEWS Modeling Environment What it is / what it’s not How it works Components, samples Installation NEWS 2 Implementation Emilio MayorgaMay.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
GNE: Global NEWS Modeling Environment
What it is / what it’s notHow it works
Components, samplesInstallation
NEWS 2 Implementation
Emilio Mayorga May [email protected] http://www.marine.rutgers.edu/globalnews/GNE
GNE Training Session 1
Content last edited August 28, 2008
Generic GNE vs. MA Scenarios Specifics
• Covering all details of Millennium Assessment (MA) Scenarios implementation is not an objective of these training sessions
• But hands-on sessions and demonstrations will be based on MA 2000 (c00) scenario
• Will use MA implementation to provide concrete examples
GNE Training Goals• Introduce & describe GNE
– Design goals, main features, code files
• Describe a standard GNE model run– Major components, files (code, configuration, and input/output
data), organization and conventions
• Describe input pre-processing and output post-processing
• Enable “model users” to set up and execute model scenario experiments– Modify input data sources used (especially for MA scenarios)– No Python coding needed
Refer to GNE User Manual for more complete details. Sessions provide more interactive training and demonstration, while manual includes extensive details about syntax, usage, installation, etc.
Training Sessions
• Session 1. Introduction and Broad OverviewThis Presentation
• Session 2. Running NEWS models with pre-existing, basin-aggregated input tablesInteractive Session, in 2 parts (2 sub-sessions)– Part 1. Model run organization, configuration files, conventions– Part 2. Model execution; changing a model run (“experiments”)
• Session 3. GIS pre-processing and model input/output post-processing
Interactive Session
Session 1 Content, Goals• Overall GNE design goals, highlights• Why Python?• Three GNE User Types
– Training focus on “Model Users” (no Python coding needed)
• GNE vs. end-to-end Global NEWS Run– What is covered by GNE vs. what is not
• Major Components– Code, configuration files, input/output files, file organization and
conventions
• GNE Model Run folder & files• GNE Model Run• GIS Pre-Processing (GNE vs. other)• Post-Processing (GNE vs. other)• Python Installation and Base GNE Setup
Overall GNE design goals, highlights
1. Dissolved models merged into a unified framework, rather than isolated sub-models; particulate models remain separate.
2. Focus on basin-scale calculations.3. For users not needing to modify model equations, most
interactions are handled via structured text configuration files. • Together with model equations in code, these files also serve as documentation
of input files used and how output was created.
4. Complete model run can be maintained in a folder structure that includes all input data, model code, constants used, configuration files, and output tables.
5. Python/NumPy. Fairly easy to learn. Model equation code implemented in an intuitive, simple array syntax.
• More complex GNE code handling input/output and overall model coordination hidden away and most users never have to see it.
6. Model run time (including GIS pre-processing) is short or trivial.
Why Python?• Free, open source, high-level but general programming
language with large user base in all fields of computing. Free = available to everyone.
• Runs on all major operating systems.• Rapidly growing acceptance in science applications,
including modeling & GIS.• Very clean, simplified syntax that makes it easy to learn,
read existing code, and become productive quickly.• NumPy “add-on” used to code the model equations;
intuitive, simple array syntax similar to Matlab.• Large library of default and optional add-ons contributed
by users, for all sorts of applications.• Preferred language chosen by ESRI for data processing
automation in ArcGIS (ArcGIS “GeoProcessing”).
Three GNE User Types
1. Model Users. Setup or change a model run using a set of text configuration files; organize input and output files in a regular, structured fashion.
Python skills not required. This training is focused on Model
Users!
2. Model Developers. Focus on translating model equations into Python GNE code using simple code tools for variable management provided by the GNE framework. Largely isolated from the “behind-the-scenes”, low-level GNE functionality that handles file read-write and basic error checking. Should be fairly comfortable programming with Python & NumPy.
3. GNE Developers. Modify any aspect of GNE functionality, including model equations.
Should be very comfortable programming with Python & NumPy.
GNE vs. end-to-end Global NEWS Run
MNPDiffuse sources
MNPPoint sources
UNHHydrology
“Raw”, HeterogeneousGIS Source files
Diffusesources
Pre-processed, homogenizedGIS Input files
Pointsources
Hydrology
Manual steps,Ad-hoc automation Direct
Global NEWSinputs
Basin-scaleInputs tables
Global NEWSoutputs
Model Outputs
GNE
Complicated variables other than basin means/sums.Eg, dam retention, FQrem, lithology class
Glo
bal N
EW
S
mod
el(s
)
GN
E G
IS
pre-
proc
essi
ngSummaries,Analysis,
Maps
Mass fluxes
Other post-processing
GNE post-processing
Major Components• Overall folder organization
– Model run code, configuration, input/output files, etc.
– Avoid blank spaces or characters other than underscores and alphanumeric characters anywhere in the folder path!
– Testing
• See User Manual for complete installation details, links to online Python/NumPy resources
Next, in Session 2 …• Session 1. Introduction and Broad Overview
This Presentation
• Session 2. Running NEWS models with pre-existing, basin-aggregated input tablesInteractive Session, in 2 parts (2 sub-sessions)– Part 1. Model run organization, configuration files, conventions– Part 2. Model execution; changing a model run (“experiments”)
We will change and execute a few model runs
• Session 3. GIS pre-processing and model input/output post-processing