1 ELIMINATING UNCERTAINTY ASSOCIATED WITH CLASSIFYING SOIL TYPES IN DISTRIBUTED HYDROLOGIC MODELING HAPUARACHCHIGE P. HAPUARACHCHI, ANTHONY S. KIEM, HIROSHI ISHIDAIRA, JUN MAGOME and KUNIYOSHI TAKEUCHI Takeuchi-Ishidaira Lab., Interdisciplinary Graduate School of Medicine and Engineering, Yamanashi University, Takeda 4-3-11, Kofu, Yamanashi 400-8511, JAPAN This paper presents a method to estimate the parameter values associated with soil characteristics in a distributed hydrologic model. The new approach accounts for all soil types in a catchment, regardless of the number of soil types present, eliminating the need for subjective soil reclassification methods and more accurately approximating the true physical characteristics of the catchment. The hydrologic model parameter values corresponding to soil type are calculated as a function of the amount (%) of basic soil types (sand, silt and clay) within each sub unit (e.g., grid, hydrological response unit or sub catchment) based on three coefficients (which represent the actual parameter value for sand, silt and clay respectively). The three coefficients can be determined based on either field measurements or by calibration (or in the case of ungauged catchments by transferring coefficient values previously determined in other similar catchments). The current study uses the new method in the BTOPMC model (block-wise use of TOPMODEL with Muskingum-Cunge flow routing method) to estimate lateral transmissivity (T 0 ). The results indicate that the performance of distributed hydrologic models can be improved using the new method and that reclassifying or grouping soil types so that the effects of soil on hydrologic processes can be modeled is a significant and unnecessary source of uncertainty that can be eliminated. INTRODUCTION The impact of soil type on the basic hydrological behaviors of a catchment is an important factor in distributed hydrologic modeling. However in any catchment, numerous different types of soil can be present making reclassification of soil type difficult and some times unrealistic. In a distributed hydrologic model, it is difficult to incorporate different parameter values for each soil type due to the large amounts of data and calculations required to estimate such a large number of parameters. Therefore in many applications, only the dominating soil types present in the catchment are identified and the minor types are incorporated into these dominating classes considering their similar properties. Thus a great deal of subjectivity and uncertainty is inevitable in these reclassification methods since it is hard to determine exactly which soil types have similar properties (e.g., is sandy clay more similar to clay or sand and what criteria are used to decide).