Distributed Hydrological Simulation System Dhsvm Of Java Implementation

With the development of hydrology, computer science and geo-information systems (GIS), the research of hydrology has entered a new era. Distributed hydrology models (DHMs) are replacing lumped ones due to their clarity in physical representation. Distribute hydrology models have particular advantages in research works on the influence of mankind activities and natural environment changes to basin water circulation, and on the occruuence and variation of regional water resource. With the advance of related technologies, the physical process represented by a DHM is becoming more and more closer to that of the real world. On the other hand, GIS provides technique support to the analysis of spatial data even in a temporal manner. The foundational data of GIS is specified by a Digital Elevation Model (DEM), which essentially provide many important parameters of the watershed. Consequently, DEM based DHM has bcome the most important trend of modern hydrology.The Distributed Hydrology-Soil-Vegetation Model (DHSVM) is an explicitly distributed hydrological model that simulates the land surface water and energy balance at the scale of a DEM. It is typically applied at high spatial resolutions on the order of 100 m for watersheds of up to 104 km2 and at sub-daily timescales for multi-year simulations. It was originally developed in the early 1990s by Mark Wigmosta while in the Washington University. Then it became popular in many universities and other scientific research institutions, with about 200 faculty users world-wide. Some researchers revised and extended the model according to their own requirments. However, the latest formal version (v3.0) was published by the Washington University in 2003. It is still written in C without graphics user interface (GUI). Fusion with other systems is also hard.This thesis describes a Java implementation of DHSVM. The major parts include physical mechanisms, system structure, functions of each part and some detailed characteristics. Other problem encountered during system implemention, such as structure and parameter adjusting, and different implementations of a particular model, are discussed in more detail. However, the work discussed in this thesis is far from a simple translation process from the C version to the Java Version. The new version has a number of advantages, including but not limited to (1) object oriented. Vegetation, soil, stream among others can be represented by objects, and physical processes of them can be clearly described by methods of them. The partition of different models/scenarios to packages and classes are often straightforward; (2) GUI provided. The generation of the configuration file can be done via a user-friendly GUI, in which online help is also provided; (3) fusion with other systems supported. For example, the data mining tool Weka can be easily incorporated to provide new implementation of some modules; (4) multithreading supported. Expending it to a parallel version, even a version running on many web browsers is not hard. In general, the work of this thesis is the foundation of future researches of our group.