How to Calculate Rainfall over a Region
If you have a large region of land, precipitation and rainfall will not be uniform. This article details how to calculate rainfall with various methods to find rainfall for a given region.
The Goal
We want to find out how much rainfall Iowa has received on a given day.
The Problem
We have data from rainfall of various weather stations across Iowa, that are not equidistant and variously dispersed around the state. We need a method that can effectively combine the various measurements of the stations to give us a singular number of rainfall for the state. We want to use observations and not use any of the re-analysis models.
Solution Overview:
- Gather all data from the various stations
- Define a method to estimate precipitation values at each grid point
- Calculate the estimated precipitation values
- Analyze and report
For now we will assume that we have a perfect dataframe containing all the stations in a region on a daily basis. Thus skipping step 1.
For step 2 there are a couple foremost methods to estimate the precipitation at each grid point: IDW method or the Kriging method.
IDW stands for Inverse Distance Weighted. This all stems from a concept that locality determines effect. A good article outlining the concept from ArcGIS
Kriging is an advanced geostatistical method that makes certain assumptions about the data. I'm going to pass on this method for now because of the nuance and complexity
Other methods here as well.
Because we are going to be using IDW we now need two additional method definitions:
- Distance function
- Grid composition to compute the distance function