No abstract

Total contributing area (TCA) is defined as upslope projection area contributing runoff to a terrain object of our interest (Erskine et al., 2006;Lindsay, 2003;Rieger, 1998). Terrain object can be a point, a pixel or a segment of terrain contour. As surrogate of runoff volume, TCA has been widely accepted as crucial terrain attribute of modeling various topographically-driven processes, such as rainfall-runoff response (

Abstract. The bias in atmospheric variables as well as that in model computation are two major causes of failures in discharge estimation. Attributing the bias in discharge estimation becomes difficult if there is a lack of qualified meteorological observations. The problem is more complicated over the mountainous area where strong orographic effects exist and with severe heterogenous geography (e.g. the Upper Tarim basin). In this study, we in the first step improved the forcing inputs using local refined precipitation dataset and glacier simulations over the upper Tarim basin. The discharge bias was then obtained by comparing the estimated discharge from an advanced land surface model (ORCHIDEE) with in-situ discharge observations. We then introduced a framework with Budyko approach which succeeded in attributing the estimated bias to possible biases from major forcing variables and model structure. The possibility of these biases was discussed by referring to many other studies with similar climatic or land surface characteristics. Results show that the water inputs (rainfall, snowfall or glacier melt) are very likely underestimated especially for the headwater catchments of the upper Yarkand and the upper Aksu, with the largest range of 43.2 % and 33.9 % respectively. Meanwhile, the potential evapotranspiration is unrealistically high over the upper Yarkand and the upper Hotan (1240.4 mm/yr and 1153.7 mm/yr respectively). The bias in actual evapotranspiration which is determined by the model strcuture is possible but not the only contributor to the discharge underestimation (overestimated by 76.1 %, 19.1 % and 105.8 % for its three headwater catchments). Based on the discharge simulation and bias analysis, we estimated the water consumption by human intervention ranging from 213.50 × 108 m/yr to 300.58 × 108 m3/yr, which is another bias source in current version of ORCHIDEE. This study succeeded in retrospecting the bias from the discharge estimation to multiple bias sources of the atmospheric variables and the model structure, although the framework needs further argumentations about its robustness, it provides a unique method for evaluating the regional water cycle and its biases with our current knowledge of observational uncertainties.

The calibration of highly parameterized hydrological models is a major computational challenge, especially for models with long run times. This challenge motivates the reconsideration of gradient‐based algorithms often overlooked for their perceived lack of robustness. Our study evaluates two Gauss‐Newton algorithms, robust Gauss‐Newton (RGN), and Levenberg‐Marquardt (PEST), and two stochastic algorithms, Shuffled Complex Evolution (SCE), and Dynamically Dimensioned Search (DDS), on a 38‐parameter SWAT model calibration problem. Algorithm performance is comprehensively assessed using trajectory plots from 100 invocations and by analyzing the distribution of estimated optima at fixed budgets of 200, 500, 1,000, 2,000, 3,000, and 5,000 objective function evaluations (model runs). Empirical results indicate that: (a) Gauss‐Newton algorithms are more likely than stochastic algorithms to locate good solutions for the budgets considered in this work, and more likely to locate satisfactory solutions when budget is tight (200–500 model runs) and (b) RGN shows the fastest initial convergence amongst the algorithms under consideration and has the highest chance of finding satisfactory solutions when the budget is tight. The results indicate that Gauss‐Newton algorithms offer an attractive choice for the calibration of highly parameterized hydrological models.