College of Social Sciences, UH Mānoa

Hawaiʻi strengthens its ability to map croplands

Posted:

More CSS News

Progress is being made on successfully mapping Hawaiʻi’s agricultural lands with high accuracy and crop specificity, thanks to a collaboration with the federal government and the research efforts of UH Mānoa professors, including Qi Chen in the Department of Geography and Environment.

The National Agricultural Statistics Service (NASS) in the U.S. Department of Agriculture announced the release of the Hawaiʻi Cropland Data Layer (HCDL), an innovative geospatial data product. The high-resolution, crop-specific dataset – the first of its kind for the Aloha State – was made publicly available in late August via NASS’s geospatial portals, CroplandCROS and AgriWatch.

NASS collaborated with UH Mānoa to develop HCDL by using Google Earth Engine and Google’s DeepMind AI-powered data. Hawaiʻi was previously a “desert” for agricultural maps, said Professor Chen in the College of Social Sciences.

He said the lack of such maps limited the ability of policymakers, land managers and researchers in Hawaiʻi to monitor crop diversity, evaluate land-use change and design programs that support food security and sustainable agriculture.

“This gap in knowledge became especially evident during the 2023 Maui wildfires, when USDA and state agencies had only limited capacity to assess the agricultural impact. Without up-to-date, field-scale crop maps, agencies were unable to quickly quantify the extent of cropland loss, identify which crops were most affected or prioritize recovery resources,” said Chen. “Instead, assessments had to be pieced together from outdated maps, secondary sources and on-the-ground reports, delaying an accurate picture of the disaster’s effect on Hawaiʻi’s agricultural sector.”

Chen said developing such maps for Hawaiʻi is particularly critical because of the state’s unique agricultural landscape. Unlike large-scale monocultures (the practice of growing a single species of plant over a large area), which are common on the U.S. continent, Hawaiʻi’s farms are often small, fragmented and characterized by diverse crops cultivated side by side.

The complexity poses both a challenge and opportunity. “While it makes mapping more difficult, it also means that accurate, high-resolution crop maps can provide transformative insights into resource allocation, irrigation planning, invasive species management and resilience to environmental change,” said Zhe Li, project co-director and geographer in the USDA.

Li added that once annual crop maps for Hawaiʻi are available, they can be integrated with real-time satellite data on weather, drought and wildfire risk to safeguard agricultural production.

“Consider a situation similar to the 2023 Maui wildfires: If high-resolution crop maps had been in place, emergency managers could have quickly overlaid fire perimeters with known crop locations to estimate economic losses and identify which producers needed the most help immediate support,” said Chen. “Beyond disaster response, the same maps could also be used proactively – by identifying cropland areas most vulnerable to drought or invasive specific, with agencies directing irrigation resources, extension services or pest management programs to the farmers who need them most.”

NASS plans to release HCDL for the 2025 crop year in February 2026.

In addition to Chen and Li, research team members include Noa Lincoln, researcher in the Department of Tropical Plant and Soil Sciences in the UH Mānoa College of Tropical Agriculture and Human Resources; Zhengwei Yang, geographer with USDA; Haonan Chen, associate professor of electrical and computer engineering at Colorado State University; and Changyong Cao, chief of NOAAʻs Satellite Calibration and Data Assimilation Branch in the Satellite Meteorology and Climatology Division.


Man with glasses at a podium presenting.
Qi Chen presented Hawai‘i Cropland Data Layer maps and introduced preliminary crop and fire/vegetation mapping at the AgriWatch Workshop.
Map landscape
Hawaiʻi Cropland Data Layer