The RE model application in an endoreic river basin of central Asia gives the evidence that the runoff absorbing processes can considerably affect hydrological response and should be incorporated into hydrological modeling. The approach is characterized in several aspects mentioned below. i) The runoff absorbing processes are reflected directly and the lateral and vertical runoff-evaporation processes are incorporated into hydrological modeling. ii) Not only river discharge but also the detail water dispersive and consumptive processes inter-river basin are represent in the RE model. That will be helpful to assessing water resources and managing groundwater and surface water wisely. The approach was employed with longer time scale because runoff-evaporation is a more sluggish and continual process relative to rainfall-runoff. This approach is suitable for water resources research whose interest is not individual precipitation event simulation. iii) Both natural processes and human dominated processes are considered in order to represent integrated hydrological cycle. The land surface transformation and the development of canal utility factor can to be represented because of the relatively longer time scale. iv) The calculation of evapotranspiration is more reliable by taking into account not only atmospheric forcing evaporation and rain-induced soil water content but also the influence of vegetation type, growth stage, groundwater table and irrigation water, management level for irrigation area and so on.
Results of the application show that the ecological valuable areas are mainly supported by the runoff abstraction in the Akesu oasis. This indicates that the status of ecosystem might be an index of runoff absorbing processes. Although recent advances in hydrological simulation in arid areas make it widely accepted that runoff absorbing processes can considerably affect hydrological response, it is difficult to represent the absorbing processes because of the knottiness to describe heterogeneity in land surface and soil type as well as the regionalized impacts of human activities. Any better knowledge we may gain about the processes of runoff-evaporation is key to understanding and predicting the hydrological response to climatic and environmental modifications and to assessing future water resources. The RE model introduced in this paper is demonstrated appropriate and efficient for typical RAA, assembling rainfall-runoff and runoff-evaporation processes and testing their applicability in semi-arid area for hydrological models is a challenge for future research on understanding hydrological cycle and managing water resources.