The accurate estimation and monitoring of carbon stored in forest and other ecosystems is a key need in carbon markets and for our carbon neutral future. Recent studies have shown forests sequester twice as much carbon as they emit which amounts to 7.6 billion tonnes of CO2 per year (Harris et al., 2021), which is about 20 times more CO2 than Australia emits annually. For this reason, we need reliable methods to measure how much sequestered carbon we have across the globe, and we need the ability to monitor this going forward and backward in time. Earth observation (EO) data in combination with cloud computing offers a solution to predicting forest carbon at scale. EO data can provide global, seamless coverage and indirectly estimate forest carbon through optical and radar satellites. EO data can also help us assess how much deforestation risk there is for current forest carbon stocks and areas which have the potential to be reforested (called carbon projects). Carbon projects often have to be stable forested ecosystems which aren’t prone to disturbance such as wildfire or anthropogenic deforestation. Open access EO data products can help us find suitable project areas with data on historical deforestation, wildfire, and distance to road.
     NGIS and Olam Food Ingredients (OFI) have been working together to develop a cost-effective method of assessing shade tree carbon of coffee and cocoa crops. Crops such as coffee and cocoa are typically grown in tropical to subtropical areas with productive and carbon dense forests. Growing conditions can vary from full sun to shade covered with various tropical tree species. Land managers from organisations such as OFI need tools to assess how much carbon is stored in these shade trees as more shade grown coffee/cocoa means more carbon stored in forests and less carbon in the atmosphere.
     This assessment of shade tree carbon is made possible due to the Global Ecosystem Dynamics Investigation (GEDI) mission (Dubayah et al., 2022; Duncanson et al., 2022). This mission generates three-dimensional LiDAR waveform data and provides information about height, density, and biomass of Earth’s forest in 25 m footprints (Figure 1).
library(dplyr)
library(ggplot2)
library(knitr)
library(plotly)
library(caret)
library(rgee)
library(RColorBrewer)
library(leaflet)
ee_Initialize(drive = TRUE, quiet = TRUE)
pal <- c("#440154", "#433982", "#30678D", "#218F8B", "#36B677", "#8ED542", "#FDE725")
sa <- ee$Geometry$BBox(-123.518, 48.45539, -123.4380, 48.49101)
gedi <- ee$ImageCollection('LARSE/GEDI/GEDI04_A_002_MONTHLY')$
filterBounds(sa)
gedi = gedi$select('agbd')$median()$clip(sa)
labels <- c("1", "2", "3", "4", "5", "6", "7")
visAGB <- list(min = 0, max = 600, palette = pal)
Map$centerObject(sa, 14)
colorbar <- leaflet::colorNumeric(palette = visAGB$palette, domain = NULL)
Map$addLayer(gedi, visAGB, name = 'gedi', ) +
Map$addLegend(visAGB, name = "AGB",
position = "bottomright")Figure 1: GEDI L4A 25 m raster footprint visualised in Google Earth Engine over the forests of British Columbia, Canada with inset picture showing the forest canopy from the ground.
Figure 2: Reporting of shade tree carbon in cocoa crops. The total plot was predicted to have 70.1 T/ha carbon. 5.7 T/ha of the total being crop carbon and 64.3 T/ha of shade tree carbon.