Use of Midlatitude Soil Moisture and Meteorological Observations to Validate Soil Moisture Simulations with Biosphere and Bucket Models

Soil moisture observations in sites with natural vegetation were made for several decades in the former Soviet Union at hundreds of stations. In this paper, the authors use data from six of these stations from different climatic regimes, along with ancillary meteorological and actinometric data, to demonstrate a method to validate soil moisture simulations with biosphere and bucket models. Some early and current general circulation models (GCMS) use bucket models for soil hydrology calculations. More recently, the Simple Biosphere Model (SiB) was developed to incorporate the effects of vegetation on fluxes of moisture, momentum, and energy at the earth's surface into soil hydrology models. Until now, the bucket and SiB have been verified by comparison with actual soil moisture data only on a limited basis. In this study, a Simplified SiB (SSIB) soil hydrology model and a 15-cm bucket model are forced by observed meteorological and actinometric data every 3 h for 6-yr simulations at the six stations. The model calculations of soil moisture are compared to observations of soil moisture, literally "ground truth," snow cover, surface albedo, and net radiation" and with each other. For three of the stations, the SSIB and 15-cm bucket models produce good simulations of seasonal cycles and interannual variations of soil moisture. For the other three stations, there are large errors in the simulations by both models. Inconsistencies in specification of field capacity may be partly responsible. There is no evidence that the SSiB simulations are superior in simulating soil moisture variations. In fact, the models are quite similar since SSiB implicitly has a bucket embedded in it. One of the main differences between the models is in the treatment of runoff due to melting snow in the spring-SSiB incorrectly puts all the snowmelt into runoff. While producing similar soil moisture simulations, the models produce very different surface latent and sensible beat fluxes, which would have large effects on GCM simulations.