Improvement of Objective Analysis of Lake Surface State in a NWP Model Using Satellite Observations
Kheyrollah Pour, H.¹; Duguay, C.R.¹; Rontu, L.²; Eerola, K.²; Kourzeneva, E.²; Fernández Prieto, D.^3
1University of Waterloo, CANADA; ²FMI, FINLAND; ³ESA, ESRIN, ITALY

Data assimilation has widely been used to solve the initial value problems at numerical weather prediction (NWP). There is a variety of users and applications of space-borne observations in NWP systems; however, this is the first time to our knowledge that a remotely-sensed Lake Water Surface Temperature (LWST) product has been tested in the pre-operational NWP environment for improvement of the weather forecast using the optimal interpolation method. Three-dimensional NWP model, High Resolution Limited Area Model (HIRLAM), experiments were performed over northern Europe for two winters (2010-2011 and 2011-2012) as part of ESA’s STSE North Hydrology project.

Results of prognostic parameterizations, based on the Freshwater lake model (FLake) implemented in HIRLAM, were compared to the analysis based on in-situ and space-borne LWST observations. The satellite observations can help to improve the description of lake surface state, especially during the ice-cover melting period. As in-situ observations of LWST are barely available for NWP applications, it is important to take space-borne obserations into account in the analysis. It was found that correct lake surface temperature may have a significant local influence on cloud cover and 2-m air temperature forecasts, although the HIRLAM standard verification against the SYNOP observations did not show significant systematic difference due to the differences in analysed lake surface state.

A challenge for NWP development is to properly combine the analysed and predicted lake surface state, using advanced data assimilation methods. Although results from this study clearly demonstrate the benefits of assimilating space-borne LWST observations into HIRLAM, and that comprehensive assimilation of LWST observations can improve NWP results, some practical issues need to be resolved in the near future before this data source can be fully adopted for operational use by NWP consortia in Europe and elsewhere around the world.