Accessible Remote Sensing Data Mining Based Dew Estimation

Dew has been considered a supplementary water resource as it constitutes an important supply in many ecosystems, especially arid and semiarid areas. Remote sensing allows large-scale surface observations, offering the possibility to estimate dew such regions. In this study, by screening combining different remote variables, we obtained well-performing monthly scale yield estimation model based on support vector machine (SVM) learning method. Using daytime nighttime land temperatures (LST), normalized difference vegetation index (NDVI), three emissivity bands (3.929–3.989 µm, 10.780–11.280 11.770–12.270 µm) inputs, simulated site-scale achieved correlation coefficient (CC) of 0.89 root mean square error (RMSE) 0.30 (mm) for training set, CC = 0.59 RMSE 0.55 test set. Applying Heihe River Basin (HRB), results showed that annual ranged from 8.83 20.28 mm/year, accounting 2.12 66.88% total precipitation, with 74.81% area having amount 16 19 mm/year. We expanded application Northwest China 5~30 mm/year 2011 2020, indicating is non-negligible part balance area. As cycle, use can provide better future assessment analysis.