Location: Hydrology and Remote Sensing LaboratoryTitle: Evaluation of satellite leaf area index in California vineyards for improving water use estimation
|KANG, Y. - US Department Of Agriculture (USDA)|
|Kustas, William - Bill|
|NIETO, H. - University Of Alcala|
|White, William - Alex|
|TORRES-RUA, A. - Utah State University|
|ALSINA, M. - E & J Gallo Winery|
|KARNELL, A. - Ben Gurion University Of Negev|
Submitted to: Irrigation Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/29/2022
Publication Date: 6/9/2022
Citation: Kang, Y., Gao, F.N., Anderson, M.C., Kustas, W.P., Nieto, H., Knipper, K.R., Yang, Y., White, W.A., Torres-Rua, A., Alsina, M., Karnell, A. 2022. Evaluation of satellite leaf area index in California vineyards for improving water use estimation. Irrigation Science. https://doi.org/10.1007/s00271-022-00798-8.
Interpretive Summary: Evapotranspiration (ET) is a critical indicator for irrigation management by measuring water loss from soil and plant. Energy balance ET models use surface temperature and Leaf Area Index (LAI) to partition evaporative fluxes between soil and plant. However, LAI estimation is subject to errors due to uncertainties from remote sensing signals and model assumptions. This paper evaluates six remotely sensed approaches for LAI estimation from Landsat and Sentinel-2 data and assesses the sensitivity of ET modeling due to the uncertainty of LAI. The paper focuses on three vineyard sites as part of the Grape Remote sensing Atmospheric Profile and Evapotranspiration eXeperiment (GRAPEX) project. Results show that the rule-based regression models trained with ground LAI measurements and satellite surface reflectance achieved high accuracy. These findings are valuable in understanding uncertainties of LAI and ET estimations from satellite remote sensing and help improving water used monitoring and irrigation scheduling.
Technical Abstract: California vineyards face a pressing need for sustainable irrigation management to optimize water use as extreme droughts become more frequent into the future. Remote sensing estimation of evapotranspiration (ET) is a direct indicator of plant water stress useful for irrigation scheduling. Many ET models use Leaf Area Index (LAI) derived from satellites to partition water fluxes, which is subject to high uncertainties. Yet, the impact of LAI uncertainties on ET remains elusive. Here we assessed six satellite-based LAI estimation approaches at 20 – 30m resolutions using ground measured LAI from four vineyards distributed across California, as part of the Grape Remote sensing Atmospheric Profile and Evapotranspiration eXperiment (GRAPEX). We evaluated the sensitivity of ET from the thermal-based two-source energy balance (TSEB) model to LAI. Results show that LAI estimated from Landsat and Sentinel-2 images based on radiative transfer modeling predicted low to medium LAI well but significantly underestimated LAI by up to 50% in highly clumped vine canopies. Rule-based regression models (Cubist) trained with ground LAI measurements and satellite surface reflectance achieved high accuracy (RMSE ~ 0.3 – 0.48, R2 ~ 0.77 – 0.93) without bias in all vineyards. TSEB ET changed proportionally to errors in LAI. But ET was much more sensitive to positive LAI bias (up to 50% changes in ET vs. 50% in LAI) than negative ones, as LAI increase elevated canopy stress and significantly reduced transpiration. Cautiously, even with minor changes in ET, errors in soil evaporation and plant transpiration were sizable as their responses to LAI were divergent. These findings call for careful consideration of satellite LAI uncertainties for ET modeling and especially for the partitioning of water loss between vine and soil or cover crop.