Abstract
Land surface variables, such as soil moisture, are among the most important components of memory for the climate system. A more accurate and long time series of land surface data is very important for real-time drought monitoring, for understanding land surface–atmosphere interaction, and for improving weather and climate prediction. Thus, the ultimate goal of the present work is to produce a long-term “land reanalysis” with 1) retrospective and 2) real-time update components that are both generated in a manner that remains temporally homogeneous throughout the record. As the first step of the above goal, the retrospective component is reported here. Specifically, a 51-yr (1948–98) set of hourly land surface meteorological forcing is produced and used to execute the Noah land surface model, all on the 1/8° grid of the North American Land Data Assimilation System (NLDAS). The surface forcing includes air temperature, air humidity, surface pressure, wind speed, and surface downward shortwave and longwave radiation, all derived from the National Centers for Environmental Prediction–National Center For Atmospheric Research (NCEP–NCAR) Global Reanalysis. Additionally, a newly improved precipitation analysis is used to provide realistic hourly precipitation forcing on the NLDAS grid. Some unique procedures are described and applied to yield retroactive forcing that is temporally homogeneous over the 51 yr at the spatial and temporal resolution, including a terrain height adjustment that accounts for the terrain differences between the global reanalysis and the NLDAS. The land model parameters and fixed fields are derived from existing high-resolution datasets of vegetation, soil, and orography. The land reanalysis output from the Noah land surface model consists of eight energy balance components and skin temperature, which are output at 3-hourly intervals, and 15 other variables (i.e., water balance components, surface state variables, etc.), which are output at daily intervals for the period of 1 January 1948 through 31 December 1998.
Using soil moisture observations throughout Illinois over 1984–98 as validation, an improvement in the simulated soil moisture (of the Noah model versus a forerunner leaky bucket model) is illustrated in terms of an improved annual cycle (much better phasing) and somewhat higher anomaly correlation for the anomalies, especially in central and southern Illinois. Nonetheless, considerable room for model improvement remains. For example, the simulated anomalies are overly uniform in the vertical compared to the observations, and some likely routes for model improvement in this aspect are proposed.
An ensemble Kalman filter dual assimilation of thermal infrared and microwave satellite observations of soil moisture into the Noah land surface model
