Research On The Mechanism And Model Of Actual Evapotranspiration From Satellite And Flux Observations

Actual evapotranspiration(ETa)is a key variable and important link between the water and energy cycle in terrestrial ecosystems.Accurate estimates of ETa are needed to address a wide range of problems such as global climate change,water balance computation,agricultural management,and hydrological modeling.In order to explore the relationship between ETa and potential evapotranspiration(ETpen)and study practicability of estimated ETa models,this study analyzed the spatial and temporal variation characteristics of ETpen and meteorological factors and their response relationship in mainland China based on this data set China Meteorological Forcing Dataset(ITPCAS)reanalysis data.And the response relationships between ETa and meteorological factors with different underlying surfaces were analyzed based on flux observation data.On this basis,the relationships between ETa and ETpen at different time scales were analyzed.Based on a generalized complementary relationship model the relationships between ETa and ETpen were quantitatively analyzed,and the relationships between evaporation proportional hypothesis and complementary correlation hypothesis at different temporal and spatial scales were unified explained.Meanwhile,the four complementary relationship models were compared and analyzed at different sites with different underlying surfaces and a typical watershed.ETa of the typical watershed was estimated by the model which had the best performance.The main results of the thesis are as follows.(1)The modified meteorological factors in the ITPCAS reanalysis data set and the estimated net surface radiation(Rn)had high accuracy and good applicability in various regions of mainland China.Based on this data set,the temperature showed an upward trend on the 91.8%of mainland China during 1979—2015 with an average change rate of 0.388?/10a.The precipitation showed an upward trend on 82.4%of the regions with an average change rate of 2.19 mm/a.The wind speed showed a downward trend on 62.1%of the regions with an average change rate of-0.062(m/s)/10a.The vapor pressure deficit showed an upward trend on 87.6%of the regions with an average change rate of 0.034 kPa/10a.Rn showed a downward trend on 70.5%of the regions with an average change rate of -1.433(MJ/m~2)/10a.(2)In mainland China,more than half of the stations and regions did not exist the phenomenon of "evaporation paradox".The ETpen showed an upward trend on 62.14%of the regions with an average change rate of 1.13 mm/a,and had a relatively high rising rate in autumn and winter.Temperature and wind speed were the main influence factors of ETpen changes(3)ETpen in mainland China was positively correlated with precipitation at multi-year scales and negatively correlated with precipitation on annual scale.ETpen was negatively correlated with precipitation in the wet season or wet regions,but not significantly correlated with precipitation in the dry season or dry regions.ETpen was negatively correlated with the humidity index of the evaporation complementary correlation hypothesis,while positively correlated with the humidity index of the Budyko hypothesis(4)Rn was the main controlling factor that affect ETa changes at the scale of 30min,while Rn and temperature were controlling factors at the daily scale.The wetter the underlying surface was in northern China,the greater ETa/Rn could be.ETa of different underlying surfaces had different responses to precipitation events in different seasons.ETa was mainly affected by underlying surface type,Rn and temperature(5)When the specific humidity index was selected,the typical watershed and flux sites were consistend with the evaporation complementary correlation hypothesis.Asymmetry was universal in evaporation complementary relationship,when ETpen estimated by the unmodified Penman formula was used.The wetter the region or time was,the stronger the positive correlation between ETa and ETpen could be.And the drier the region or time was,the stronger the negative correlation between ETa and ETpen could be(6)The performance of generalized complementary relationship models were better than that of traditional complementary relationship model.By comparing multiple complementary relationship models,the sigmoid generalized complementary relationship model had obtained good verification results at flux station with different underlying surfaces.The sigmoid generalized complementary relationship model was used to estimate ETa in a typical watershed(source region of the Yellow River)which had a good estimation precision.The ETa in source region of the Yellow River showed an upward trend on 96.6%of the regions with an average change rate of 0.45 mm/a.On the annual scale ETa changes were mainly affected by temperature and Rn,while on the monthly scale ETa changes were mainly affected by R_n.