Characterisation of Water and Heat Fluxes of Typical Crops and Simulation of Evapotranspiration in Saline‐Alkali Soil of the Yellow River Delta, China

ABSTRACT

The investigation of water and heat flux variation patterns in saline‐alkali land is significant due to the distinctive characteristics that affect crop growth, and surface energy flux and evapotranspiration are two key factors affecting saline‐alkali land. Surface energy fluxes and evapotranspiration of three crops (wheat, maize, and soybean) in saline‐alkali soils were observed using an EC (eddy covariance) system. The energy balance closure of the three crops was evaluated at the daily scale with regression slopes of 0.82 for wheat, 0.83 for maize, and 0.65 for soybean. During the growing season, wheat, maize, and soybean exhibited average LE (latent heat) to Rn (net radiation) ratios of 0.66, 0.55, and 0.65, respectively. Notably, LE dominated the consumption of Rn. The correlation analysis showed that the three crops in saline‐alkali soil had the highest correlation with Rn and photosynthetic photon flux density (PPFD) and a negative correlation with humidity (RH). Notably, crops in saline‐alkali soil exhibited more pronounced nocturnal evapotranspiration (ET) variations in the middle and late growth stages compared to other dryland cropping systems. Based on the observed ET data and meteorological factors, this paper constructed ET prediction models for three crops in saline‐alkali soil at 30‐min and daily scales using back propagation neural network (BP), random forest (RF), and k‐neighbourhood (KNN). BP exhibits better model performance. In most cases, the results showed that the best model performance of wheat, maize, and soybeans at the 30‐min scale were R² = 0.812, RMSE = 0.0449 mm; R² = 0.683, RMSE = 0.0858 mm; R² = 0.802, RMSE = 0.0672 mm. The work in this study contributes to the understanding of the changing patterns of water–heat fluxes in crops in saline‐alkali soil and enables prediction of crop evapotranspiration in saline soils.