15 May 2024

Artificial intelligence accurately measures soil evapotranspiration

,

Researchers at the University of Illinois have used artificial intelligence to predict missing data. Algorithm reduces margin of error compared to traditional measurements

by Matteo Cavallito

 

Artificial intelligence can help estimate moisture levels from remote soil surveys. This is supported by a study by the University of Illinois Urbana Champaign. The survey makes an important contribution in measuring a key variable. The total amount of evapotranspiration, or the process by which water moves from the soil to the atmosphere, is subtracted from the volume of precipitation to determine the available water balance. Assessing this phenomenon, however, scientists observe, can be complicated.

Linking ground-based data with satellite data

“Ground-based ET estimates capture the local fluxes of water transferred to the atmosphere but are limited in scale,” explained Jeongho Han, lead author of the research, in a statement released by the University of Illinois. “In contrast, satellite data provide ET information on a global scale. Still, they are often incomplete due to clouds or sensor malfunction, and the satellite cycle over an area may require several days.”

Thus, the goal of the work was to develop a system that could “to predict missing data and to generate daily continuous ET data that accounts for the dynamics of land use and atmospheric air movement.”

In detail, the authors developed a model called Dynamic Land Cover Evapotranspiration Model Algorithm (DyLEMa) that can predict evapotranspiration data through trained seasonal machine learning systems. The researchers collaborated with the National Center for Supercomputing Applications (NCSA) and the Illinois Campus Cluster Program (ICCP) to process the information and train the models. With the idea of making the collected data available to other researchers.

Artificial intelligence reduces the range of error

The authors evaluated the performance of DyLEMa in a confined 30-meter x 30-meter area in the Illinois territory using data from the past two decades made available by NASA, the US Geological Survey and the National Oceanic and Atmospheric Administration. The researchers, in particular, tested the accuracy of the model by comparing the results with existing data.

The algorithm’s predictions “reduced the average PBIAS error from + 31 % to −7% compared to existing US Geological Survey evapotranspiration products,” the research states.

New applications

According to the study, DyLEMa could be used to evaluate different contexts and datasets by reducing the margin of uncertainty in estimating evapotranspiration. In fact, the system would be able to evaluate different variables including land use and, in the case of agricultural land, different crops including corn and soybeans. The model can integrate factors such as precipitation, temperature, moisture, solar radiation, vegetation stage and soil properties.

In this way, the model would fit well with the assessment of agricultural land in which crops change rapidly.

Finally, the algorithm provides important data for assessing soil erosion (the study, not coincidentally, is part of a larger research project on the topic by the U.S. Department of Agriculture). “Evapotranspiration controls the soil moisture content and vice versa, which impacts surface processes such as runoff and erosion,” explains Maria Chu, professor and co-author of the research. “Our next step is to integrate our data in a distributed hydrological model for better estimation of soil erosion.”