trigger-model

Flood trigger model

This repository contains scripts for the different steps involved in Impact based forecasting (IBF) model development The scripts avilable in this repo grouped into the fowllowing catgories. Geo spatial data needed for running the scripts are stored in 510geonode server. For running the scripts please check Geo_setting.R for name of variables for retriving data from 510geonode server.

1. Understanding the risk

Impacts from natural disasters in the historical period will be analyised to understand the existing risk, both on economies and people's lives. we developed a scipt which is used to asses flood/drought risk based on historical data.

Input

• Admin boundaries and impact data are uploaded on 510 geonode server
• Geo_setting.R file is updated with layer names for these datasets

Output

• The outcome of this analysis is a risk map, which is based on historical impact data

Ethiopia Risk Analysis Example

Kenya Risk Analysis Example

Uganda Risk Analysis Example

2. Impact hazard catalogue

A script which link impact (which was analyzed in understanding the risk part) with Hazard. Hydro-Met data will be used as a proxy for discribing hazard.

Input

• Admin boundaries and impact data are uploaded on 510 geonode server
• Geo_setting.R file is updated with layer names for these datasets
• River discharge/Rainfall data

Output

• An Impact Hazard Catalog (A csv file with a daily record of flood impact and associated Hazard)
• plots for each events showing Impact and Hazard link(For EAP)
• The outcome of this analysis is a hazard-impact catalog, which will be an input for the trigger table(next step)

Hazard river discharge based on GLOFAS

Example of impact-hazard analysis for a single event(hazard is based on glofas data) Ethiopia Example of impact-hazard analysis for a single event(hazard is based on glofas data) Kenya

Hazard rainfall based on NAM hstorical data

Example of impact-hazard analysis for a single event(hazard is based on NAM rainfall data) Ethiopia

3. Trigger model development

Based on impact-hazard catalog developed in the above step a scriptis developed to generate a trigger table the table is generated for a set of predefined trigger ranges and lead times, which can be modified by the user. For generating the first version of the trigger table we used lead time 3 days and 7 days. Five values for trigger threshold, Q10,Q50,Q70,Q90 and Q95, were defined to generate the first version of the trigger table. For calculating the trigger table we made the following assumptions

• a time window of two weeks was selected when checking Glofas peak values against reported impact incidents(this is to consider uncertainty in impact data reporting date)
• after a trigger value is passed, any other event in a time window of 3 weeks is considered the same event as the first one
• the above assumptions can be adjusted easily if needed

Here we build a trigger model based on data from Impact-hazard catalogue

4. Visualization of trigger model

easy visualization of the IBF model results for descion makers

• a shyny app to easily visualize impact hazard catalog shyny app for visualizing impact-hazard catalog data) Ethiopia

Notes

To run catchment extractor script addtional data is needed - this can be found in the following link https://rodekruis.sharepoint.com/:f:/s/510-CRAVK-510/EhG4P2uRQRZKjZiJlo7YMYwBs5sqYxzcHmElbF4GtCGF6Q?e=g2mdMV

GLOFAS data

• Uganda river discharge, 2000-2019
• Kenya river discharge, 2000-2019
• Old files: https://rodekruis.sharepoint.com/sites/510-CRAVK-510/Gedeelde%20%20documenten/FbF%20-%20Mali%20-%20RPII/Trigger%20methodology/glofas2.nc.tar