Search Results

You are looking at 1 - 10 of 13 items for

  • Author or Editor: Michael H. Freilich x
  • All content x
Clear All Modify Search
Michael H. Freilich

Abstract

A statistically consistent and physically realistic approach for validating vector magnitude measurements is developed, based on a model for random measurement noise that explicitly satisfies a nonnegativity constraint for all “noisy” vector magnitude measurements. Numerical and analytic approximations are used to quantify the nonlinear functional dependence of sample conditional means on true values and component noise magnitudes. In particular, it is shown analytically that random component errors will result in overall vector magnitude biases. A simple nonlinear regression of measured sample conditional mean vector magnitudes (calculated from traditional collocated data) against Monte Carlo simulation results is proposed for determining both deterministic trends and random errors in the data to be validated. The approach is demonstrated using Seasat and ERS-1 scatterometer measurements and collocated buoy data. The approach accounts well for the observed qualitative features of the collocated datasets and yields realistic values of random component error magnitudes and deterministic gain and offset for each dataset. An apparent systematic insensitivity of scatterometers at low wind speeds is shown to be a consequence of random component speed errors if it is assumed that the comparison buoy measurements are error free.

Full access
Dudley B. Chelton and Michael H. Freilich

Abstract

Wind measurements by the National Aeronautics and Space Administration (NASA) scatterometer (NSCAT) and the SeaWinds scatterometer on the NASA QuikSCAT satellite are compared with buoy observations to establish that the accuracies of both scatterometers are essentially the same. The scatterometer measurement errors are best characterized in terms of random component errors, which are about 0.75 and 1.5 m s−1 for the along-wind and crosswind components, respectively.

The NSCAT and QuikSCAT datasets provide a consistent baseline from which recent changes in the accuracies of 10-m wind analyses from the European Centre for Medium-Range Weather Forecasts (ECMWF) and the U.S. National Centers for Environmental Prediction (NCEP) operational numerical weather prediction (NWP) models are assessed from consideration of three time periods: September 1996–June 1997, August 1999–July 2000, and February 2002–January 2003. These correspond, respectively, to the 9.5-month duration of the NSCAT mission, the first 12 months of the QuikSCAT mission, and the first year after both ECMWF and NCEP began assimilating QuikSCAT observations. There were large improvements in the accuracies of both NWP models between the 1997 and 2000 time periods. Though modest in comparison, there were further improvements in 2002, at least partly attributable to the assimilation of QuikSCAT observations in both models.

There is no evidence of bias in the 10-m wind speeds in the NCEP model. The 10-m wind speeds in the ECMWF model, however, are shown to be biased low by about 0.4 m s−1. While it is difficult to eliminate systematic errors this small, a bias of 0.4 m s−1 corresponds to a typical wind stress bias of more than 10%. This wind stress bias increases to nearly 20% if atmospheric stability effects are not taken into account. Biases of these magnitudes will result in significant systematic errors in ocean general circulation models that are forced by ECMWF winds.

Full access
Michael H. Freilich and Barry A. Vanhoff

Abstract

One year of collocated, rain-free normalized backscatter cross-section measurements from the Tropical Rainfall Mapping Mission (TRMM) precipitation radar (PR) and wind speed estimates from the TRMM Microwave Imager are used to construct fully empirical model functions relating cross section to wind speed for incidence angles from 0° (nadir) to 18°. With the exception of a ∼1.9-dB offset, the TRMM PR model function at nadir compares well with TOPEX and Geosat results. For the first time using spaceborne data, least squares fits of the TRMM PR model functions to the geometric optics scattering formulation allow direct solution for the magnitudes and wind speed dependencies of Ku-band effective nadir reflectivity and effective mean square slope [s(u)]. Effective reflectivity is found to decrease slightly with increasing wind speed above 3.5 m s–1. A logarithmic dependence of s(u) is found for wind speeds between 1 and 10 m s–1. Both linear and two-branch logarithmic dependencies (with a slope break at 10 m s–1) are excellent and statistically indistinguishable parameterizations for s(u) for wind speeds between 5 and 19 m s–1. Calculations using the Elfouhaily et al. wave model suggest that the effective Ku-band s(u) corresponds to a cutoff wavelength of 3 to 6 times the radar wavelength for wind speeds from 5 to 25 m s–1; at lower wind speeds, the cutoff wavelength increases rapidly consistent with earlier observations.

Full access
Michael G. Schlax, Dudley B. Chelton, and Michael H. Freilich

Abstract

Sampling patterns and sampling errors from various scatterometer datasets are examined. Four single and two tandem scatterometer mission scenarios are considered. The single scatterometer missions are ERS (with a single, narrow swath), NSCAT and ASCAT (dual swaths), and QuikSCAT (a single, broad swath obtained from the SeaWinds instrument). The two tandem scenarios are combinations of the broad-swath SeaWinds scatterometer with ASCAT and QuikSCAT. The dense, nearly uniform distribution of measurements within swaths, combined with the relatively sparse, nonuniform placement of the swaths themselves create complicated space–time sampling patterns. The temporal sampling of all of the missions is characterized by bursts of closely spaced samples separated by longer gaps and is highly variable in both latitude and longitude. Sampling errors are quantified by the expected squared bias of particular linear estimates of component winds. Modifications to a previous method that allow more efficient expected squared bias calculations are presented and applied. Sampling errors depend strongly on both the details of the temporal sampling of each mission and the assumed temporal scales of variability in the wind field but are relatively insensitive to different spatial scales of variability. With the exception of ERS, all of the scatterometer scenarios can be used to make low-resolution (3° and 12 days) wind component maps with errors at or below the 1 m s−1 level. Only datasets from the broad-swath and tandem mission scenarios can be used for higher-resolution maps with similar levels of error, emphasizing the importance of the improved spatial and temporal coverage of those missions. A brief discussion of measurement errors concludes that sampling error is generally the dominant term in the overall error budget for maps constructed from scatterometer datasets.

Full access
Dudley B. Chelton, Michael H. Freilich, and Jerry R. Johnson

Abstract

Ambiguity in wind direction has long been an impediment to applications of wind observations from the Seasat scatterometer (SASS). Three months of unambiguous global SASS vector winds (7 July-10 October 1978) have recently become available from the Goddard Space Flight Center(GSFC) Laboratory for Atmospheric Sciences. The directional ambiguities were removed objectively through the use of an atmospheric general circulation model. Elimination of the directional ambiguities greatly enhances the utility of the SASS wind observations. Atmospheric and oceanographic applications of the GSFC vector wind data are given in two companion papers.

The GSFC SASS vector wind dataset has potential utility for many applications, so it is useful to assess the data quality. In this paper, the GSFC data are compared with 14 days of SASS vector winds from the Jet Propulsion Laboratory for which the directional ambiguities were subjectively removed by trained meteorologists using pattern recognition techniques. Both methods are shown to be statistically very similar over the 14-day period. The two methods chose the same solution about 73% of the time and nearest-neighbor solutions about 20% of the time. Reasons for discrepancies are examined and it is concluded that the methods tend to disagree most often in regions of low wind speed and/or highly variable wind direction. The directional differences are generally small and random so that there are no significant differences between spatially and temporally averaged wind fields constructed from the two datasets.

Full access
Dudley B. Chelton, Michael H. Freilich, and Steven K. Esbensen

Abstract

Measurements of near-surface winds by the NASA scatterometer (NSCAT) from October 1996 through June 1997 are analyzed to investigate the three major wind jets along the Pacific coast of Central America that blow over the Gulfs of Tehuantepec, Papagayo, and Panama. Each jet is easily identifiable as locally intense offshore winds in the lee of low-elevation gaps through the Sierra Madre mountain range. The jets have relatively narrow cross-stream width but often extend several hundred kilometers or more into the Pacific. The Tehuantepec and Papagayo jets sometimes merge with the northeast trade winds of the Pacific.

The Tehuantepec jet was highly energetic with characteristic timescales of about 2 days. Events were triggered by high pressures associated with cold surges into the Gulf of Mexico that originated over the Great Plains of North America. The Papagayo and Panama jets were much more persistent than the Tehuantepec jets. The winds at both of these lower-latitude locations exhibited a strong seasonal variation with almost exclusively offshore flow from late November 1996 through late May 1997 and periods of onshore flow in October and November during the late stages of the 1996 Central American monsoon season. Superimposed on this low-frequency seasonal variation were events with characteristic timescales of a few days.

Based on NSCAT data, the spatial and temporal evolution of major wind events is described in detail for three representative case studies. In December 1996, the jets developed sequentially from north to south, consistent with the notion that wind events in the two lower-latitude jets are associated with cold-air outbreaks that trigger the Tehuantepec jet a day or so earlier. In November 1996 and March 1997, the Papagayo and Panama jets were strongly influenced by tropical phenomena that had little apparent association with the Tehuantepec jet. These latter two case studies, together with the distinction between the statistical characteristics of the three jets, suggest that the Papagayo and Panama jets are predominantly controlled by a mechanism that is very different from the across-gap pressure gradients associated with high pressure systems of midlatitude origin that control the Tehuantepec jet.

Full access
Michael H. Freilich, Hongbo Qi, and R. Scott Dunbar

Abstract

Calculation of vector winds from spaceborne fan-beam scatterometers requires that backscatter measurements from different antennas be relatively calibrated to high accuracy. A method is developed to perform postlaunch antenna calibration using global mean ocean backscatter measurements in conjunction with estimates of the statistical distribution of near-surface wind velocity and the model function relating backscatter to winds. Substantial analytic simplifications result from assuming that the wind speed and azimuth distributions are separable and that the upwind–downwind asymmetry term in the model function is small. The analytic model allows quantitative examination of the sensitivity of the technique to errors in the approximate wind distributions and empirical model function. The approach and its assumptions are tested using 13 months of ERS-1 backscatter data, surface wind estimates from two operational weather analyses, and three empirical C-band model functions. It is shown that the ERS-1 antennas are relatively calibrated to within 0.2 dB, which is consistent with other published results obtained using ground receiving stations and Amazon forest data. The results are nearly insensitive to realistic errors in the estimated wind velocity distributions and model function. Analysis suggests that the ocean antenna calibration technique should be accurate to about 0.2 dB using as little as 3 weeks of scatterometer data.

Full access
Dudley B. Chelton, Michael H. Freilich, and Steven K. Esbensen

Abstract

Satellite estimates of winds at 10 m above the sea surface by the NASA scatterometer (NSCAT) during the 9-month period October 1996–June 1997 are analyzed to investigate the correlations between the three major wind jets along the Pacific coast of Central America and their relationships to the wind and pressure fields in the Inter-American Seas and eastern tropical Pacific. Comparisons with sea level pressure confirm the conventional view that Tehuantepec wind variations are driven by pressure variations in the Gulf of Mexico associated with North American cold-air outbreaks. The three jets sometimes developed sequentially from north to south. Statistically, however, the Papagayo and Panama jets were poorly correlated with variations of the Tehuantepec jet over the NSCAT observational period. The Papagayo and Panama jets were significantly correlated with each other and were coupled to coherent variations of the trade winds extending from the Caribbean Sea to the eastern tropical Pacific.

The detailed structures of the wind fields within the three jets are examined to infer dynamical balances within the jets. After leaving the coast, the northerly Tehuantepec and Panama jets turn anticyclonically toward the west in manners that are consistent with jets that are inertially balanced at the coast and become progressively more geostrophically balanced with increasing distance from the coast. There is no evidence of anticyclonic turning of the easterly Papagayo jet, suggesting that the winds may remain in approximate geostrophic balance through the gap over the Nicaraguan lake district.

NSCAT observations are compared with operational analyses by ECMWF to investigate the detailed structures of the wind fields over the Gulfs of Tehuantepec, Papagayo, and Panama. Systematic differences between the NSCAT observations and the ECMWF analyses of the divergent off-axis fanning of all three jets suggest that there may be systematic errors in parameterizations of boundary layer processes in the ECMWF “first-guess” fields in these data-sparse regions.

Full access
Dudley B. Chelton, Alberto M. Mestas-Nuñez, and Michael H. Freilich

Abstract

Three months of vector wind observations from the Seasat-A satellite scatterometer (SASS) are used to construct gridded fields of monthly average wind stress and wind stress curl over the global ocean. These fields are examined to identify features either poorly resolved or not present in wind stress fields constructed from conventional data. Particular attention is focused on the spatial structures in the high southern latitudes and the tropical regions. The SASS wind stress fields are compared globally with the Hellerman and Rosenstein climatological monthly average surface wind stress fields and with monthly averages of wind stress computed from contemporaneous 1000 mb wind analyses produced by the National Meterological Center.

The potential for satellite scatterometry improving the present knowledge of the global wind stress field and making important contributions to ocean modeling is demonstrated by comparison of global maps of the Sverdrup circulation computed from 3-month average SASS and Hellerman and Rosenstein climatological wind stress curl fields. Except for the region south of about 35°S, the two estimates of Sverdrup circulation are generally very similar. The transports of the western boundary currents calculated from the climatological wind stress fields are consistently higher because the drag coefficient used by Hellerman and Rosenstein is too large by approximately 19%. In the Antarctic Circumpolar Current region, the difference between the two estimates of zonal transport is very large, even disagreeing in direction in the region south of Australia and New Zealand. Historical hydrographic data suggest a tendency, albeit less extensive, for the westward Sverdrup transport indicated by the SASS data, suggesting the possibility of deficiencies in the climatological wind stress fields.

Full access
David Halpern, Michael H. Freilich, and Robert A. Weller

Abstract

The European Centre for Medium-Range Weather Forecasts (ECMWF) and Institut Francais Pour la Recherche et l’Exploitation de la Mer European Remote-Sensing Satellite, named IFR2, surface wind velocity data products are compared during July 1995 over the Arabian Sea. Substantial differences were found. The central positions of the maximum isotach were separated by 450 km, and the ECMWF maximum isotach was 2 m s−1 higher than that of IFR2. IFR2 wind components contained about 10 times more variance than ECMWF winds for horizontal distances from 50 to 250 km. Along the 8.5°N southern boundary of the Arabian Sea, ECMWF southward Ekman transport was higher than that of IFR2 by an amount that could be observed with current measurements. The ECMWF and IFR2 difference in upward transport of water into the Ekman layer, computed from wind stress curl, was large enough to measure.

Full access