AbstractsAstronomy & Space Science

Exploiting ground-based measurements of the Global Positioning System for numerical weather prediction

by Reima Eresmaa




Institution: University of Helsinki
Department: Department of Physical Sciences, Division of atmospheric sciences; Finnish Meteorological Institute
Year: 2007
Keywords: meteorologia
Record ID: 1142790
Full text PDF: http://hdl.handle.net/10138/23109


Abstract

Data assimilation provides an initial atmospheric state, called the analysis, for Numerical Weather Prediction (NWP). This analysis consists of pressure, temperature, wind, and humidity on a three-dimensional NWP model grid. Data assimilation blends meteorological observations with the NWP model in a statistically optimal way. The objective of this thesis is to describe methodological development carried out in order to allow data assimilation of ground-based measurements of the Global Positioning System (GPS) into the High Resolution Limited Area Model (HIRLAM) NWP system. Geodetic processing produces observations of tropospheric delay. These observations can be processed either for vertical columns at each GPS receiver station, or for the individual propagation paths of the microwave signals. These alternative processing methods result in Zenith Total Delay (ZTD) and Slant Delay (SD) observations, respectively. ZTD and SD observations are of use in the analysis of atmospheric humidity. A method is introduced for estimation of the horizontal error covariance of ZTD observations. The method makes use of observation minus model background (OmB) sequences of ZTD and conventional observations. It is demonstrated that the ZTD observation error covariance is relatively large in station separations shorter than 200 km, but non-zero covariances also appear at considerably larger station separations. The relatively low density of radiosonde observing stations limits the ability of the proposed estimation method to resolve the shortest length-scales of error covariance. SD observations are shown to contain a statistically significant signal on the asymmetry of the atmospheric humidity field. However, the asymmetric component of SD is found to be nearly always smaller than the standard deviation of the SD observation error. SD observation modelling is described in detail, and other issues relating to SD data assimilation are also discussed. These include the determination of error statistics, the tuning of observation quality control and allowing the taking into account of local observation error correlation. The experiments made show that the data assimilation system is able to retrieve the asymmetric information content of hypothetical SD observations at a single receiver station. Moreover, the impact of real SD observations on humidity analysis is comparable to that of other observing systems. Satelliittipaikannusjärjestelmät, kuten GPS- ja GALILEO-järjestelmät, luovat pohjan nykyaikaiselle paikannukselle. Väitöskirjassa tarkastellaan, miten paikannussignaaleja voidaan hyödyntää myös säähavaintoina sääennustusmalleissa. Väitöskirjatyö luo edellytykset ilmakehän kolmiulotteisen vesihöyryjakauman entistä tarkempaan määrittämiseen ja ennustamiseen. Väitöskirjassa kehitetyillä menetelmillä on merkitystä mm. pilvisyyden ja sateen ennustamisessa yhteispohjoismaista HIRLAM-ennustusmallia käytettäessä. Esitettyjen menetelmien avulla on lisäksi mahdollista vähentää satelliittipaikannukseen liittyvää mittauskohinaa, mikä on…