| Land surface and overlying atmosphere are strongly coupled with each other.The exchange of water and energy between them affects the atmosphere conditions and land surface hydrologic processes.Thus,the study of land-atmosphere feedback is of great significance for improving atmospheric and hydrologic simulation abilities.Traditional hydrologic modeling approaches regard the land-surface hydrologic system as an independent and open system,and focus on responses of the hydrologic system to external disturbances,while ignoring the further influences of the local land-atmosphere feedback on the land surface hydrologic system.Physically based climate models are capable of modeling complex land-atmosphere interactions,but are often involved in the concerns about uncertainties of modeling results and huge computational burden.Therefore,this study aims to identify the feedback relationships between key variables of land and atmosphere systems from observational data,simplifying the complex land-atmosphere interactions into statistical relationships,and construct a flexible and efficient land-atmosphere two-way coupling framework based on this.In this way,the hydrologic modeling considering land-atmosphere feedbacks can be realized,and thus helps to improve the understanding of hydrologic cycle and provide theoretical basis and technical support for the scientific utilization of water resources.The main contents and results are summarized as follows:(1)Investigating the statistical characteristics of the feedback relationships between land(soil moisture,evapotranspiration)and atmosphere(precipitation,temperature)variables.A “hot spot” region,state of Illinois,USA,is selected as the study area.Based on gauge observations,the feedback effects of soil moisture and evapotranspiration on the next-day precipitation(transition probabilities and amount)and temperature(daily maximum and minimum)are analyzed through Generalized Linear Model(GLM).First,key impact factors for precipitation and temperature are determined by stepwise regression,which are revealed to include not only seasonal cycle of climate system,the autocorrelation of climate variables and the crosscorrelation between variables in parallel or delayed forms on daily timescale,but also the feedback effects from land surface hydrologic variables(soil moisture and evapotranspiration).Then,climate models excluding and including land-surface feedback effects are constructed respectively,and the decrease of AIC in the latter model indicates that the introduction of land surface variables improves model fitting.Meanwhile,fitting results show that soil moisture and evapotranspiration have positive effects on precipitation triggering and amount,while have negative effects on daily maximum temperature.In addition,the analysis of fitting errors reveals decent model fitting in different spatial and temporal scales.At last,the simulation results of the GLMs with and without feedback effects are compared to the historical data,which proves that the former has better simulation ability.(2)Constructing a land-atmosphere two-way coupling model,which consists of a weather generator,and a land surface hydrologic model.First,the Variable Infiltration Capacity(VIC)hydrologic model is built based on soil and vegetation information in the study region.The model running procedure has been modified from “time-before-space” to “space-before-time” to meet the requirement of real-time feedback process.Second,a GLM-based weather generator(WG)is constructed according to the land-atmosphere feedback relationships revealed in the previous chapter.In order to consider the spatial dependence between grid cells,the simulation of precipitation is carried out by using the joint distribution of precipitation occurrence and amount.Moreover,due to the minimum requirement of climate variables of the VIC model,an independent wind speed model has been added into the WG.Therefore,the WG and the VIC model are linked to each other through precipitation,temperature,soil moisture and evapotranspiration,realizing the two-way coupling of the two models.Finally,the modeling performance of the WG-VIC model has been evaluated by historical observations.The results show that the coupled model is able to capture the temporal and spatial patterns of precipitation,temperature and soil moisture.(3)Application of the WG-VIC model in scenarios of land use/land cover(LULC)and climate changes and its comparison to one-way coupling scheme.The LULC change simulation is achieved by modifying the vegetation parameters in the VIC according to future LULC scenarios in Illinois.The simulation results show that compared with the traditional one-way coupling scheme,the land-atmosphere feedback effect will enlarge the local precipitation and surface water resources while decrease the air temperature,but the overall difference is not large.The simulation of climate change is achieved by modifying the parameters in the WG according to two Representative Concentration Pathway(RCP)climate scenarios.The simulation results show that future climate change will lead to significant increases in precipitation and temperature,while one-way coupling scheme reveals slight decreases in the soil moisture and surface water resources.The change patterns under two climate scenarios are similar,while the differences lie in the extent.The two-way coupling scheme shows that under RCP 4.5 scenario,the existence of land-atmosphere feedback has a buffering effect on the reduction of precipitation and runoff,and a certain amplifying effect on the rise of temperature,but the overall difference is not large.However,under RCP 8.5 scenario,the existence of land-atmosphere feedback may accelerate the regional water cycle in the land-atmosphere system and make the region maintained in a relatively humid condition.The application study of WG-VIC two-way coupling model reveals significant differences with one-way coupling approach which is traditionally used in hydrologic simulation. |
PDF Full Text Request