Satellite-based crop coefficient and regional water use estimates for Hawaiian sugarcane

•We estimated crop coefficient (Kc) and evapotranspiration (ET) for sugarcane in Maui.•The satellite NDVI was able to predict canopy ground cover (fc).•Crop coefficient (Kc) of sugarcane can be estimated by the satellite-based canopy crop cover.•The results were sensitive to the selection of ground weather station data and choice of reference ET.

Water availability is a major limiting factor for sustainable production of potential biofuel crops in Maui, Hawaii, USA. It is essential to improve regional, near-real time estimates of crop water use to facilitate optimal water management. Satellite remote-sensing offers multiple methods to estimate water management, however, most approaches assume a spatially homogeneous regional meteorology, which does not apply in Hawaii due to strong orographically-induced variations in climate. Canopy ground cover (fc) and spectral reflectance were measured over two sugarcane fields on the satellite Landsat 7 overpass dates. A strong linear relationship between satellite-based Normalized Difference Vegetation Index (NDVI) values and fc was found (R2 = 0.97). Canopy ground cover was also found to be highly correlated to crop coefficient (Kc) which was calculated from reference evapotranspiration (ET0) and sugarcane crop evapotranspiration (ETc) observations from the Eddy Covariance towers. Spatial ET0 for the entire sugarcane plantation was determined from an automated weather station network previously installed by the commercial grower of the plantation. Regressions were used to convert the satellite NDVI to Kc maps, and then the satellite-based ETc maps were created from the satellite-based Kc maps and ET0 values computed and interpolated from the weather station network. To compare with the satellite-based crop coefficient/reference ET method, ETc and Kc were also estimated using a satellite-based surface energy balance technique, the METRIC model. These two satellite-based methods were correlated fairly well (R2 = 0.84 for Kc and 0.81 for ETc). Compared to the METRIC model, the satellite-based crop coefficient/reference ET method may be more practicable for estimating crop water use in Maui, where ground-based ET0 measurements are available through the on-site automated weather station network with right choice of reference ET equation.