Abstract:

Accurately quantifying soil evaporation on sandy land is of great significance for developing of plantation construction for restoring the ecology of the Hobq Desert. In this study， lysimeter was used to measure soil evaporation of an artificial forest in the Yinkensha Forest Farm in the Hobq Desert. However， due to the influence of external factors， the data observed by the lysimeter would have abnormal values. The Adaptive Window and Adaptive Threshold method （AWAT） was used to process the lysimeter data. Meteorological variables were measured synchronously to assess their influences on soil evaporation. The following results were obtained： on an hourly scale， the coefficient of determination between the precipitation captured by the lysimeter and the precipitation obtained by the rain gauge was 0.93， and the average error and the root mean square error were -0.00078 mm and 0.11 mm， respectively. The probability of detection， frequency of hit， and Heidke’s index score were 0.75， 0.9， and 0.8 respectively； On the daily scale， the coefficient of determination， average error and root mean square error were 0.99， -0.02 mm， and 0.66 mm， respectively. The probability of detection， frequency of hit， and Heidke’s index score were 0.94， 0.85， and 0.85， respectively. The above statistics suggest the AWAT method had achieved in processing abnormal data from the lysimeter. During the study period， the soil evaporation of an artificial forest farm was 93.87 mm， mainly concentrated in July-September， with a maximum daily evaporation of 16.5 mm； precipitation was 264.88 mm， also concentrated in July-September， with a maximum daily precipitation of about 35 mm. The soil evaporation of sandy plantations accounted for 35.4% of precipitation. The main meteorological factors affecting soil evaporation in desert artificial vegetation areas were precipitation and solar radiation， and wind speed had little effect on soil evaporation.

Key words: Hobq Desert, lysimeter, AWAT, soil evaporation, meteorological factors