This paper examines the application of the Kalman filter for assimilation of satellite soil moisture measurement data into the SL-AV global numerical weather prediction (NWP) model. This technique allows to consider soil moisture data in areas with available satellite observations. Single-assimilation numerical experiments based on the Kalman filter revealed a reduction of errors in the initial surface layer soil moisture data.
The present paper presents characteristics of correlation between soil moisture observations obtained from a satellite and direct observations during the warm period of 2011 and 2012 in the area of the Tunka Basin. The factors influencing the relationship are considered. It is shown that the updated satellite data on moisture of the upper soil layer and those of direct observations at a depth of 15 cm have a satisfactory relationship.
The work is devoted to the assessment of the possibility of using satellite data to determine soil moisture. The direct satellite observations were compared with the direct observations at the stations. Pearson correlation coefficient and the relative errors were calculated. The comparison of the data from the direct measurements of the soil moisture and satellite measurements showed that for the 52% of the stations the correlation coefficients exceeded 0.5.
This paper presents two data filtration methods. These methods are used for filtration of satellite soil moisture measurement data. A comparison with in-situ soil moisture measurement data shows an improvement in data quality after application of the filters. First results of satellite data assimilation with a global model of numerical weather forecasting are given.
The topsoil moisture defined at meteorological stations and the one defined with the use of MetOp satellite has been compared in the paper. The comparison is made with the measurements at stations located on the USA territory and included in the observation network FLUXNET-AMERIFLUX and Soil Climate Analysis Network (SCAN). The stations are located in different climatic zones, defined according to the vegetation type as per the IGBP classification. The research period is 2007–2012. The satellite observation data are corrected to the in-situ measurements and measurement accuracy is assessed. The assessment has been done for different types of underlying terrains. A good agreement with the real data of in-situ moisture measurements has been shown, which allows to use the satellite soil moisture measurement data in data assimilation systems for numerical weather prediction models.
A high-resolution mesoscale meteorological model TSU-NM3 has been presented for use in forecasting and investigating the weather phenomena and ground air quality over a limited urbanized area and over a major industrial centre or road junction. To solve mesoscale model equations an efficient explicit-implicit differencing method of the second order of approximation has been developed and supercomputer-oriented. The calculation results are well correlated with the measurements made for the City of Tomsk.