Flow Routing in a Cratered Landscape: 2. Model Calibration for Pleistocene and Modern Lakes and Rivers of the U.s. Great Basin

Introduction: Geomorphic features such as valley networks, fans, and deltas indicate that there were fluvial processes acting on Martian surface some time in the past [1-6]. While geomorphic and mineralogical evidence for an early “wet Mars” has strengthened with recent Mars missions, the question remains as to the magnitude and duration of environmental conditions that could support surface liquid required to create observed channels and other fluvial features. The Great Basin as an analog of Martian terrain: Many terrestrial geomorphic processes have been seen as analogs for former Martian conditions. The Great Basin (GB) province in the southwestern U.S. (including the Basin and Range and Mojave regions) has many similarities with Mars, thus was chosen as the study site to validate and calibrate our flow routing and lake simulation model [7]. Like Mars, the GB consists of many enclosed basins and lakes. At present it supports an extremely dry environment with an evidence of wetter conditions in the past. Moreover, the GB has been strongly shaped by fluvial erosion and paleolacustrine deposits and shorelines indicate that the climate has fluctuated greatly during the late Quaternary [8-10]. Application of the model to this region has several benefits: 1) it allows testing of model water balance and routing, 2) it demonstrates sensitivity of lake levels and drainage network integration to variation in climatic parameters, 3) it can be incorporated into the General Circulation Models (GCMs) to model the lake-atmosphere interactions, and 4) though GB Pleistocene climate has been studied extensively, work on spatially explicit flow routing is limited. Furthermore, the GB has modest vegetation so that its influence on hydrological cycle is minimal. Data sources: Producing a hydrological model requires spatially explicit estimates of runoff and evaporation, which are in turn functions of precipitation, temperature, latitude, and elevation. Model validation in turn requires comparative data on the spatial distribution of modern and paleolakes. Relevant data were collected for California, Idaho, Nevada, Oregon, Utah, and Wyoming and reduced to the GB region (34-44° N, 109-122° W). Modern environment. Many of the data used for this study were available via websites. The digital elevation data and mean annual precipitation data were obtained from the Shuttle Radar Topography Mission (SRTM) and National Resources Conservation Services (NRCS), respectively. Both data files were converted to ASCII surfer grid format with reduced resolution of 1000 m per pixel. The National Climatic Data Center (NCDC) provides mean annual temperature (MAT) data along with the deviation form the historical average for each year for 318 stations that are within the GB. For each station, the MAT, geographic coordinates, and elevation were recorded. Microsoft Excel was used to conduct a multiple regression analysis between the MAT and latitude (lat), longitude (lon), and elevation (elev) in meters, which had resulted in a following relationship with regression R of 0.87: 137.721 1.556 0.00263 0.117 MAT lat elev lon = − − + . Lake evaporation rate is hard to estimate owing to difficultness of accurately measuring its controlling parameters like wind speed, atmospheric moisture content, and cloud cover [11]. A simpler method for estimating evaporation rate was suggested by [12], using a graph of evaporation against elevation for two different latitudes for Nevada. Using this graph as a basis, we estimated the mean annual evaporation rate, E, as a function of latitude and elevation: 4.976 0.0744 0.00062 E lat elev = − − . Similarly, mean annual runoff (Q) was expressed as a function of mean annual precipitation (P: in m yr) and MAT (F), using another graph provided by [12]: