Benchmark and partial validation testing of the FLASH computer code, Version 3.0

This document presents methods and results of benchmark testing (i.e., code-to-code comparisons) and partial validation testing (i.e., tests which compare field data to the computer generated solutions) of the FLASH computer code, Version 3.0, which were conducted to determine if the code is ready for performance assessment studies of the Radioactive Waste Management Complex. Three test problems are presented that were designed to check computational efficiency, accuracy of the numerical algorithms, and the capability of the code to simulate diverse hydrological conditions. These test problems were designed to specifically test the code's ability to simulate, (a) seasonal infiltration in response to meteorological conditions, (b) changing watertable elevations due to a transient areal source of water, (i.e., influx from spreading basins), and (c) infiltration into fractured basalt as a result of seasonal water in drainage ditches. The FLASH simulations generally compared well with the benchmark codes, indicating good stability and acceptable computational efficiency while simulating a wide range of conditions. The code appears operational for modeling both unsaturated and saturated flow in fractured, heterogeneous porous media. However, the code failed to converge when a unsaturated to saturated transition occurred. Consequently, the code should not be used when this condition occurs or is expected to occur, i.e. when perched water is present or when infiltration rates exceed the saturated conductivity of the soil