Satellite remote sensing of meteorological parameters for global numerical weather prediction

Abstract

Remote sensing of meteorological parameters helps to provide the initial conditions for numerical weather prediction (NWP). Desired fields include those of temperature, moisture, winds, clouds, and surface properties. For high horizontal resolution and global coverage, satellite data are an unrivaled source of information. The basic form of this information is satellite sensor‐incident, wavelength‐dependent radiance (or equivalently, brightness temperature). The process of retrieving meteorological information from the observed radiances involves solving an inverse problem. Often the inverse problem is ill posed or poorly conditioned. A variety of methods have been used to make the inverse problem more tractable: in essence, some a priori information is required. The inverse problem is the opposite of the calculation of radiances, given the relevant meteorological profiles, surface conditions, and sensor geometry. This so‐called forward problem is an integral part of physical retrieval methods which are now generally favored over conventional statistical retrieval methods. Physical retrieval methods, in practice, actually rely to a certain degree on a priori statistics. Temperature retrievals, especially in the southern hemisphere, are of proven importance in NWP involving global scale models. Vertical temperature profiles are obtained from observed radiances in a set of emission bands with varying optical depths. In comparison to conventional radiosonde data, current temperature retrievals have coarse vertical resolution. There is also some evidence that retrieved synoptic features are somewhat smoothed horizontally in spite of the high horizontal resolution of the temperature retrievals. The usefulness of other geophysical parameters for NWP is not well established and is largely untested, although a great many retrieval methods have been proposed or developed for a variety of potentially interesting parameters. The reason for this is twofold: these data are expected to have the most impact on mesoscale modeling, but mesoscale analysis systems are still relatively primitive. Moisture variables—that is, specific humidity, clouds, and precipitation—are retrievable and are potentially very useful. It is theoretically possible to retrieve specific humidity profiles by using methods analogous to those used to retrieve temperatures. However, results to date with available sensors have not been wholly satisfactory. Other moisture parameters are somewhat easier to retrieve; for example, cloud cover may be obtained by a variety of techniques from imagery in visible and/or infrared spectral regions, and vertically integrated water vapor is well correlated with emissions at microwave frequencies, especially over the ocean. Unfortunately, these parameters are not readily assimilated by a global NWP model, and special methods are required. A variety of other meteorological parameters which may be used for NWP are retrievable. These include winds at the surface from microwave sensors and cloud drift winds aloft, as well as o...

Atmospheric physiography and meteoronomical modelling: the future role of geographers in understanding climate

“…Dickinson 1987). The most serious conceptual difficulty hampering development of a four-dimensional, global-scale satellite data assimilation scheme (Isaacs et al 1986) is, without doubt, the problem of establishing GCM gridscale characteristics from pixel-scale observations. Wetzel (1987) shows that the recognition of spatial variability of soil moisture is essential for modelling evapotranspiration over land surfaces.…”

The potential contribution of satellite remote sensing to the understanding of arid lands processes

“…Clearly, those variables that have less effect on the measurements cannot be retrieved reliably from observed variations of F t. The problem is said to be ill-conditioned with respect to these variables: the inversion procedure may assign a value to all variables, but the reliability of the estimates may be so low as to be useless for some of these variables. These issues have been discussed by Isaacs et al (1986) in the context of numerical weather prediction.…”

Synergistic Use of MODIS and AIRS in a Variational Retrieval of Cloud Parameters

“…Many schemes have been proposed for estimating cloud parameters from multispectral radiance observations (Isaacs et al 1986;Wielicki and Coakley 1981;Susskind et al 1987;Eyre and Menzel 1989). One important method using passive remote sensing data for obtaining the altitude of mid-and upper-level clouds, especially transmissive clouds, is the CO 2 -slicing technique (Chahine 1974;Smith et al 1974;Smith and Platt 1978;Menzel et al 1983;Menzel et al 1992).…”