README.md

In impact-initiatives-geospatial/easyrgee: easy rgee - wrappers for `rgee` functions

easyrgee

The goal of easyrgee is to …

Installation

You can install the development version of easyrgee from GitHub with:

# install.packages("devtools")
devtools::install_github("impact-initiatives-geospatial/easyrgee")
remotes::install_github("impact-initiatives-geospatial/easyrgee"")

Note

The package contains mainly convenience wrappers for the rgee package. The rgee package requires a GEE account.

Much of the functionality of this package is slowly being integrated into the exploreRGEE package.

Examples

Zonal stats

This is a basic example which shows you how to solve a common problem: extracting zonal statistics of an ee$ImageCollection and plotting

First we load a few packages, initialize earth engine, and load in an internal data set which contains 3 admin polygons in NE Nigeria.

library(easyrgee)
library(rgee)
library(ggplot2)

ee_Initialize(quiet=T)

# 3 admin boundaries in NE NGA
dat<- easyrgee::adm1_ne_nga

Next we load in a MODIS NDVI ee$ImageCollection for 2016

modis_ndvi <- ee$ImageCollection("MODIS/006/MOD13Q1")$select("NDVI" )$filterDate("2016-01-01","2016-12-31")

Finally we perform the aggregation/reduction with ee_extract_long.

ee_extract_long is a tool that can be used to extract values or zonal statistics over both Images and Image collections. It is a wrapper around rgee::ee_extract. ee_extract_long uses a couple of server-side processes and client-side regex manipulations to help standardize the output while pivoting rgee::ee_extract output from wide to long format. The result is a data.frame ready for processing/plotting.

median_ndvi<- easyrgee::ee_extract_long(ic = modis_ndvi,
                              sf = dat,
                              sf_col = "ADM1_EN",
                              scale = 250,
                              reducer = "median")

Now we can easily use ggplot to visualize the results

median_ndvi |> 
  ggplot(aes(x=date, y= value, color=ADM1_EN))+
  geom_line(lwd=1)+
  scale_color_manual(name="State (NGA)" ,values =c("#ee5859","#DBD6C7", "#818183") )+
  theme_minimal()

Identify Start of Rainy Season

Start time of rainy season can have implications on crop harvest which can then impact well-being/livelihoods. Therefore, we have made a function to help identify a date which can be considered start of rainy season.

Lets first look at cumulative rainfall

Let’s load in CHIRPS 2016 daily rainfall data and then use the ee_accumulate_band_ic function to create a cumulative precipitation ee$ImageCollection which stores pixel-wise accumulation of values with each consecutive image containing the sum of it’s value with all previous Images

chirps <-  ee$ImageCollection("UCSB-CHG/CHIRPS/DAILY")$
  filterDate("2016-01-01","2016-12-31")

precip_cumulative<- easyrgee::ee_accumulate_band_ic(ic = chirps, band = "precipitation")

Next we again use ee_extract_long to extract the daily median values of 3 different geographic zones

median_cumulative_rainfall<- easyrgee::ee_extract_long(ic = precip_cumulative,sf = dat,sf_col = "ADM1_EN",scale = 5000,reducer = "median")

use ggplot to visualize rainfall accumulation curves by region

median_cumulative_rainfall |> 
  ggplot(aes(x=date, y= value, color=ADM1_EN))+
  geom_line(lwd=1)+
  scale_color_manual(name="State (NGA)" ,values =c("#ee5859","#DBD6C7", "#818183") )+
  theme_minimal()

From the above plot, let’s set 20 mm as a good threshold to use as the start of rainy season

With our threshold we can use ee_accum_reduce_to_doy which

1. Creates a cumulative precipitation ee$ImageCollection as above
2. Finds the first day the user-defined threshold was met for each pixel.
3. Returns an ee$Image ranging from 0-365 (days of year)

using the chirps 2016 daily rainfall data let’s use ee_accum_reduce_to_doy to return a day of year (doy) ee$Image where the pixel values represent the first day the cumulative rainfall surpassed our set threshold 20 mm

precip_threshold_doy<- easyrgee::ee_accum_reduce_to_doy(ic = chirps,thresh = 20,band = "precipitation")

We can use rgee::Map functions to visualize the image on a leaflet map (in this GitHub README the map is not interactive although it is in the R-session)

#visualize
Map$setCenter(zoom = 3)
Map$addLayer(precip_threshold_doy,
             visParams = list(min=0, max=365, palette='#9ecae1,#ffffff,#ffeda0,#feb24c,#f03b20'))

impact-initiatives-geospatial/easyrgee documentation built on March 26, 2022, 10:42 p.m.