Skip to content

Frequently Asked Questions

What is ERM?
ERM is an experimental tool written in Google Earth Engine (GEE) platform to monitor extreme rainfall that could trigger a flood, and the impact to population and cropland.
What is the main input for ERM?
GPM IMERG for near real-time rainfall monitoring and NOAA GFS for rainfall forecast.
JRC Global Surface Water to develop historical flood occurrence data by months.
JRC Global Human Settlement Layer for calculating population affected.
MODIS Annual Land Cover for calculating cropland affected.
What makes ERM unique compared to the competitors?
ERM has a simple output (yes or no / flood or no flood).
Analysis of critical rainfall (threshold) is conducted by pixels by months, in area with spatial resolution 0.1deg x 0.1deg ~ 10km x 10km.
What can ERM do?
ERM are able to inform where is the estimated location of extreme rainfall and its impact to population and cropland in the last 5-days and forecast up to 5-days ahead based on selected date.
Rainfall with extreme categories are well detected through ERM.
What is ERM limitation:
ERM is an experimental application, sometimes the results are accurate (flood detected) and sometimes inaccurate (flood undetected). Sorry about that!
Many factors cause the results to be inaccurate, one of them is the rainfall forecast data quality. Other factor is below.
What ERM can not do?
ERM only able monitor or predict flood caused by extreme rainfall in a location where extreme rainfall occurred. And unable to calculate flood caused by runoff accumulation from other areas.
What is difference between near real-time (NRT) and forecast (FCT)
NRT inform about the latest condition, 1 to 5 days in the past as of selected date. While the FCT inform the situation that will happen in the 1 - 5 days ahead as of selected date.
How to read ERM output?

Every computation/single process on ERM will produce 4 outputs:

  • (1) rainfall accumulation,
  • (2) rainfall exceeding the threshold,
  • (3) Likelihood of flooding, and
  • (4) flood alert. All are available for both near real-time and forecast.

Rainfall accumulation is the total rainfall for day(s) of simulation selected as of the selected date. The classification value is following the legend on sidebar.

Rainfall exceeding the threshold categorized into 4 class:

  • (1) #ffffcc Moderate, exceeding percentile 50.
  • (2) #a1dab4 Heavy, exceeding percentile 80.
  • (3) #41b6c4 Intense, exceeding percentile 90.
  • (4) #225ea8 Extreme, exceeding percentile 96. See The Threshold

Likelihood of flooding is based in Historical Flood Occurrence data, generated using JRC Global Surface Water and combine with maximum rainfall by months. See Focal Linear Regression. Categorized into 3 class:

  • (1) #f7fcb9 Low, has probability of flooding less than 60%.
  • (2) #addd8e Moderate, has probability of flooding between 60-80%.
  • (3) #31a354 High, has probability of flooding greater than 80% or we can say this area are frequently flooded.

Flood alert visualized into 4 categories: Green, Yellow, Orange and Red. See below explanation.

What do the colors on flood alert represent?

#97D700 Green, is a condition when one of the following conditions is met:

  • Rainfall exceeding percentile 50 (Moderate rainfall) and the likelihood is less than 60% (Low).
  • Rainfall exceeding percentile 50 (Moderate rainfall) and the likelihood is 60-80% (Moderate).
  • Rainfall exceeding percentile 80 (Heavy rainfall) and the likelihood is less than 60% (Low).

#FFEDA0 Yellow, is categorized into 3 class (Alert class 1, 2 and 3) and a condition when one of the following conditions is met:

  • (Alert 1) Rainfall exceeding percentile 50 (Moderate rainfall) and the likelihood is greater than 80% (High).
  • (Alert 2) Rainfall exceeding percentile 80 (Heavy rainfall) and the likelihood is 60-80% (Moderate).
  • (Alert 3) Rainfall exceeding percentile 90 (Intense rainfall) and the likelihood is less than 60% (Low).

#FEB24C Orange, is categorized into 3 class (Alert class 4, 5 and 6) and a condition when one of the following conditions is met:

  • (Alert 4) Rainfall exceeding percentile 80 (Heavy rainfall) and the likelihood is greater than 80% (High).
  • (Alert 5) Rainfall exceeding percentile 90 (Intense rainfall) and the likelihood is 60-80% (Moderate).
  • (Alert 6) Rainfall exceeding percentile 96 (Extreme rainfall) and the likelihood is less than 60% (Low).

#F03B20 Red, is categorized into 3 class (Alert class 7, 8 and 9) and a condition when one of the following conditions is met:

  • (Alert 7) Rainfall exceeding percentile 90 (Intense rainfall) and the likelihood is greater than 80% (High).
  • (Alert 8) Rainfall exceeding percentile 96 (Extreme rainfall) and the likelihood is 60-80% (Moderate).
  • (Alert 9) Rainfall exceeding percentile 96 (Extreme rainfall) and the likelihood is greater than 80% (High).

If an areas experience a Green alert, we assume won't have any significant impact.

ERM use Alert 1 to 9, and included in the impact calculation.

See Matrix for final alert

How to read impact calculation?

Both impact analysis using population and cropland are categorized into 3 class:

  • #FFEDA0 Yellow (Low),
  • #FEB24C Orange (Moderate) and
  • #F03B20 Red (High impact).

These 3 class are consistence with flood alert class.

Is ERM mobile friendly?
No, ERM not optimized for mobile devices. Please use your computer or laptop to access it.
What is ERM future plan?
See development plan in the left sidebar.