Integration of satellite imagery and in situ soil moisture data for estimating irrigation water requirements

Remote sensing images provide a reliable approach to monitor crop physiological development and can be used to assess crop evapotranspiration (ETc) and irrigation water requirements (IWR). This study compared the suitability of multispectral images acquired from Unmanned Aerial Vehicles (UAV-MSI), PlanetScope, and Sentinel-2A & 2B satellite platforms for estimating ETc. It integrated this ETc data with in situ soil moisture data to estimate IWR of field gown tomato crops (Lycopersicum esculentum) in southeastern Canada. The experimental field was divided into three (3) blocks, and irrigation scheduling consisted of 100, 80, and 60% of soil's field capacity, corresponding to three irrigation regimes. Plants were selected from each of the three blocks, through a systematic grid sampling approach. The sampled plants were georeferenced and identified in the images. Normalized difference vegetation indices (NDVI) obtained from the remote sensing platforms were evaluated for estimating the crop consumptive coefficient and ETc. The ETc predicted from satellite images were compared with estimates of ETc obtained from the FAO 56 Penman-Monteith module of the AquaCrop model. ETc maps from Sentinel-2 were combined with soil moisture data to predict IWR. The results indicate a significant difference in average NDVI values obtained from the UAV-MSI (0.87 +/- 0.03) and satellite platforms (0.71 +/- 0.03 and 0.82 +/- 0.05, for PlanetScope and Sentinel-2, respectively), which suggests that the UAV-MSI overestimated the field NDVI values. There was a good agreement between Kc and NDVI values extracted from satellite images, with R-2 = 0.98, p < 0.001, for Sentinel-2 and R-2 = 0.78, p < 0.001, for PlanetScope. ETc values estimated from Sentinel-2 satellite platform were closely corroborated with AquaCrop model (with R-2 = 0.94; p < 0.01), which shows the suitability of Sentinel-2 imagery for assessing crop canopy cover and IWR at the field scale. The amount of irrigation water that the grower applied using micro-drip irrigation system (342 and 416 mm in 2017 and 2018 growing seasons, respectively) exceeded the estimated IWR (165 and 199 mm in 2017 and 2018 growing seasons, respectively), which suggests that the field was over-irrigated. This study has shown the practicality of integrating soil moisture measurements and remotely sensed crop parameters for mapping actual irrigation requirements. It indicates a significant progress towards the development of a near real-time approach for supporting precision irrigation.

Automated summary

Remote sensing images are a reliable approach for monitoring crop physiological development and estimating crop evapotranspiration and irrigation water requirements. This study compared the suitability of different remote sensing platforms for estimating crop water requirements and found that Sentinel-2 imagery was highly suitable for assessing crop canopy cover and irrigation water requirements at the field scale. The study also revealed that the field was over-irrigated, with the amount of irrigation water applied exceeding the estimated irrigation requirements.