Find areas at risk of flooding
Historical Flood Occurrence
The data presents the occurrence of surface water for the selected period. A 100% water occurrence means that all data points in the collection are identified of water (permanent water), 50% implies a location that is inundated half of the time.
Locations with a low surface water occurrence have only been inundated for a short period in the whole period, likely caused by flood events.
HFO was done based on script from Dr. Ate Poortinga, its written in his blog post https://mygeoblog.com/2017/10/20/historical-flood-occurrence/
GEE script to calculate HFO: https://code.earthengine.google.com/08f671f8be1190b61a1a6175caf5b429
Focal Linear Regression
To get areas at risk, we need to relate the maximum rainfall and historical flood occurrence.
Then construct pair data of flood occurrence MEAN by month and maximum of rainfall by month for the period 2000-2019 for all IMERG grids, by performing linear regression between every 5×5 pixels of the two rasters, we would like to create a new raster of SLOPE (a) and INTERCEPT (b).
Each pixel of the SLOPE and INTERCEPT will hold the regression slope and intercept value obtained from linear regression of the corresponding 5×5 pixels that surround that pixel.
We develop R script to do focal regression computation: https://github.com/wfpidn/Focal-Regression but the result still not perfect. So we did it manually in Excel spreadsheet using standard formula.
Will it trigger a flood?
As an example, it is found that for pixel at (6,2) equation that can estimate the probability of flood event is:
gi = 0.01888 * Xi - 4.0219
Using the above formula, with the availability of rainfall forecast for the X-days, the gi can be estimated and translated into probability of flood.
For example, if the forecast at X-days reaches 300 mm (categorized as Extreme rainfall, exceeding Percentile 96) then:
gt = 0.01888 * 300 - 4.0219
gt = 1.6421
The corresponding probability
pt = 1/(1+exp(-1.6421)
pt = 0.837820
categorized as High likelihood
Then ALERT category 9 will release for this event.
Combination from Extreme Rainfall and High Likelihood
Matrix for final alert
