A new high-resolution map of Hawaiʻi has been released to reveal trends in what crops farmers are growing — and can potentially be used to assess and even predict wildfires.
The map was made as part of a collaboration between the University of Hawaiʻi and the U.S. Department of Agriculture, which provided $270,000 for the project. Researchers used Google Earth Engine and Google’s DeepMind, an artificial intelligence system, to do the work.
The map, called the Hawaiʻi Cropland Data Layer, is free and publicly available through the USDA’s National Agricultural Statistics Service website.
A previous satellite-based map showing how farm land is used in Hawaiʻi was released in 2020, but the new HCDL can be made more quickly. Now changes in land and crops can be monitored annually.
“We are able to do this for the first time in Hawaiʻi and using a fully automated computer-based approach. … So, we'd be able to do it every year in a highly automated way,” said Qi Chen, a professor at UH Mānoa’s Department of Geography and Environment and the project’s director.
The 2020 map was made more manually, so it can’t be done annually — meaning it can be difficult to monitor closely the changes in how farm land is used.
Chen also said the new map can also be more specific, which is particularly helpful in Hawaiʻi because of all the specialty crops grown on the islands.
“In the old map there's a big category called ‘diversified crops,’ but in our new maps, we try to break down that big category as long as we’re able to map it. For example, if there's a big farm, there's an area of corn, there's papaya or other trees, and previously maybe they were lumped into one class,” he said.
Work is ongoing to make the map even more functional. The map has data available for 2023 and 2024, but can potentially show changes in crops and farm land over the course of a single year.
The high-resolution map can be used for non-agricultural purposes, too. It has data on forests and shrub coverage, water and even urban infrastructure.
Chen said the project was in part a response to the Maui fires in 2023, when the landscape changed dramatically and rapidly.
The ultimate goal is to integrate weather data and fire modeling so that the map can be used to assess wildfire damage, predict an area’s risk for wildfires and even track the movement of a fire in real time.
“Beyond disaster response, the same maps could also be used proactively — by identifying cropland areas most vulnerable to drought or invasive species, with agencies directing irrigation resources, extension services or pest management programs to the farmers who need them most,” Chen said in a statement.
