Introduction

The land surface interface is a major component of the climate system since it provides coupling between the atmosphere and the land biosphere, hydrology. The land surface is characterized by spatial heterogeneity that spans a wide range of scales. Several land surface models were developed to describe the land¨Catmosphere water and energy exchanges, such as bucket model (Manabe, 1969), Biosphere-Atmosphere Transfer Scheme (BATS) (Dickinson et al., 1986), the Simple Biosphere (SiB) model (Sellers et al., 1986), and the Bare Essentials of Surface Transfer (BEST) (Desborough and Pitman, 1998). These models emphasized vertical structure, representing the land surface as on or two tiers of vegetation (canopy or groundcover, or both). One of the main shortcomings of these schemes is that they do not capture the pronounced heterogeneity in the earth's land surface. This heterogeneity affects the surface energy and water budgets, as well as the land-atmosphere exchanges of momentum, heat, water through a number of nonlinear processes. Distributed representation of spatial information and physical description of land biosphere and hydrology processes are necessary because of the highly nonlinear processes and the variability of the spatial heterogeneity. The resolution of present-day general circulation models (GCMs) is still too coarse to explicitly capture the effects of surface heterogeneity, which therefore needs to be parameterized within the framework of complex and nonlinear land surface process schemes. A realistic representation of subgrid variability would constitute a significant improvement for a land surface model (Koster and Suarez, 1992).

Investigations of subgrid scale variability associated with terrain, soil and vegetation heterogeneities have been motivated by many studies. Milly and Eagleson (1988) found the potential for a serious underestimation of surface runoff if the areal variability of precipitation associated with storms of various scales and types is ignored. Entekhabi and Eagleson (1989) used analytic distributions of rainfall and soil moisture conditions to examine the sensitivity of runoff, bare soil evaporation efficiency, and transpiration efficiency to soil type and climatic forcing. Avissar and Pielke (1989) suggested a parameterization of the subgrid-scale forcing of heterogeneous land surfaces for atmospheric numerical models and found that spatial heterogeneity in vegetation could have significant effects on temperature and precipitation. Pitman et al. (1990) used a surface hydrology model driven by meteorology simulated by a GCM to investigate the influence of the subgrid distribution of precipitation on the surface water balance. Their results indicated that improving the realism of the areal distribution of precipitation could alter the balance between runoff and evaporation. Seth et al. (1994) divided one GCM grid into several subgrids to study the effects of subgrid-scale vegetation and climate specification on surface fluxes and hydrology. They showed that energy partitioning at the surface, surface stress, and runoff could all be significantly affected by subgrid variability. Ghan et al. (1997) presented a preliminary evaluation of the relative importance of subgrid variations in parameters related to surface hydrology. They found that subgrid variability in summertime precipitation would increase runoff, and subgrid variations in vegetation and soil properties would increase surface runoff and reduce evapotranspiration. Giorgi (1997a); Giorgi (1997b) described a theoretical framework for the representation of surface heterogeneity within complex biophysical surface schemes for use in climate models and assess its sensitivity to relevant parameters.

Giorgi and Avissar (1997) presented a review of methodologies for the representation of land surface subgrid scale heterogeneity effects and grouped the effects of surface heterogeneity into two categories: "aggregation" and "dynamical" effects. Subgrid scale aggregation has been shown to affect the simulated surface latent and sensible heat fluxes, snowpack, and dynamics of soil moisture and runoff. Dynamical heterogeneity effects are associated with microscale and mesoscale circulations induced by heterogeneity surfaces. Models of aggregation effects attempt to calculate the contribution of different subgrid scale surface types to the grid box average energy and water budgets and surface-atmosphere exchanges. Such models have been based on discrete approaches, whereby heterogeneity is described in terms of a finite number of subgrid "tiles" or "patches" and on continuous approaches, in which heterogeneity is described in terms of probability density functions. The precipitation and soil characteristics variability is investigated based on continuous approaches with probability density functions in many researches (Yeh and Eltahir, 2005; Zeng et al., 2002; Entekhabi and Eagleson, 1989; Gao and Sorooshian, 1994; Liang and Xie, 2001). The land use and vegetation cover subgrid variability is represented based on discrete approaches by several studies (Koster and Suarez, 1992; Leung and Ghan, 1998). Koster and Suarez (1992) considered two conceptually different strategies: the "mixture" and "mosaic" strategies, for dealing with subgrid variability in vegetation cover. The mixture strategy assumes that the different vegetation types are effectively mixed homogeneously throughout the grid square, so that the atmospheric interacts only with a set of near-surface atmospheric conditions pertaining to the mixture. With the mosaic strategy, the different vegetation types in a grid square are assumed to be geographically distinct. The different types are viewed as separate tiles of a grid square mosaic, and each tile interacts with the atmosphere independently. The effective differences between the strategies are small over a wider range of condition. In particular, the strategies are effectively equivalent when the transpiration resistances of the different vegetation types are of the same order of magnitude.

Although the subgrid variability on natural factors, such as precipitation, soil infiltration capacity and vegetation cover, is studied by many researchers, there are little studies investigating the subgrid variability caused by human activities. Döll and Siebert (2002) modeled the global irrigation water requirements under present-day climate conditions and found annual irrigation water requirement in hot semi-arid regions can be more than 1000 mm. Boucher et al. (2004) concluded that human activity has a direct influence on the water vapour concentration through irrigation, and estimated a global mean radiative forcing up to 0.1 Wm^-2, and a surface cooling of up to 0.8 K over irrigation area. Gordon et al. (2004) showed that deforestation is as large a driving force as irrigation in terms of changes in the hydrological cycle. Haddeland et al. (Available online 23 November 2005) reported an irrigation scheme in a macroscale hydrological simulation and evaluated the effects of irrigation on the water and energy balances of the Colorado and Mekong river basins. These researches indicate that the subgrid variability caused by human activities is potentially important to affect the surface water and energy balance. However, there are few complete researches to describe the effects of both the subgrid variability of natural factor and human activities on hydrological simulation. Especially, few researches were done to represent the influence of subgrid variability on large scale distributed hydrological pattern inside a large river basin.

Among the subgrid heterogeneities affecting hydrological processes, we account for two heterogeneities, precipitation heterogeneity and the heterogeneity on irrigation redistributing runoff, representing the natural subgrid variability and the variability caused by human activities, respectively. The precipitation heterogeneity is represented with a simple spatial exponentially distribution. An irrigation scheme based on simulated soil moisture and available water is developed to represent subgrid variability related to irrigation. The objective of the study is to analyze the effects of anthropogenic heterogeneity on the water and energy balances of a large-scale basin in semi-arid river basin. Comparisons are made between effects of natural heterogeneity and anthropogenic heterogeneity.