Collaborative Research: Generation of Improved Land-surface Data and an Assessment of its Impact on Mesoscale Predictions

改进近地表变量数值模拟的一个主要问题是模型预测对关键地表参数准确性的敏感性。 特别是，近地表量的预测，如2米的温度和相对湿度和10米的风强烈影响的土地表面的状态，是降水。 预测这些参数的能力对于各种各样的人类活动都很重要，从规划户外和周末活动到运输路线和节能。 植被特征，如植被覆盖度（FVEG）和叶面积指数（LAI）可以说是最重要的土地表面参数，需要准确地定义。 然而，目前定义这些参数的做法过于简单，通常只基于气候学。 然而，植被对降雨量的每日变化做出反应，远非静态的。 PI最近的研究表明，指定FVEG和LAI，在高空间和时间分辨率，在一个国家的最先进的耦合大气-陆地表面modeling systems.This research is a unique collaboration of expertise between three institutions in the key areas of mesoscale atmosphere modeling，land surface modeling and the generation of real time，high resolution，satellite derived land surface parameters. 这项工作的智力价值包括：1）根据近地面变量的日常预报对各种陆面模式进行评估和改进，2）第一个四维同化系统，它既有标准观测网络数据，也有来自俄克拉荷马州Mesonet的原始通量数据，以及3）根据极地轨道卫星的每日数据开发一个自动化知识系统，以提供陆地表面变量的每日更新。 这项研究的重点区域是大平原，特别是在2003年的生长季节（3月至10月）。这项研究的更广泛的影响是最终改善近地表变量，如温度和湿度的短期预测，从而影响电力负荷，空气质量和对流天气预报，所有这些都具有显着的经济意义。

A major problem area for improved numerical modeling of near surface variables is the sensitivity of model predictions to the accuracy of key land surface parameters. In particular, the forecasts of near surface quantities such as 2 m temperatures and relative humidity and 10 m winds are influenced strongly by the state of the land surface, as is precipitation. The ability to predict these parameters is important to a wide variety of human activities, ranging from planning outdoor and weekend activities to transportation routing and energy conservation. Vegetation characteristics such as fractional vegetation coverage (FVEG) and leaf area index (LAI) arguably are the most important land surface parameters that need to be defined accurately. However, present practices for defining these parameters are overly simplistic and typically are based only on climatology. Yet vegetation responds to daily variations in rainfall and is far from static. Recent studies by the PIs demonstrated potentially large impacts from specifying both FVEG and LAI, at high spatial and temporal resolution, in a state-of-the-art coupled atmosphere-land surface modeling system.This research is a unique collaboration of expertise between three institutions in the key areas of mesoscale atmospheric modeling, land surface modeling and the generation of real-time, high resolution, satellite-derived land surface parameters. The intellectual merit of this work consist of: 1) an evaluation and improvement of various land surface models from the routine daily predictions of near surface variables, 2) the first four-dimensional assimilation system of both standard observational network data and raw flux data from the Oklahoma Mesonet, and 3) the development of an automated knowledge-based system based upon daily polar-orbiting satellite data to provide daily updates of land surface variables. The focus region for this study is the Great Plains, especially during the growing season of 2003 (March to October).The broader impacts of this research are eventual improvements to short-range predictions of near surface variables such as temperature and moisture, thereby affecting power load, air quality, and convective weather forecasting, all of which have significant economic implications.