Search

Experiments were conducted to assess the impact of SEASAT-A scatterometer (SASS) data on high resolution model predictions of the QE2 storm and to explore different approaches for utilizing SASS data. For all of the experiments, the model used is the limited-area model with a horizontal resolution of 100 km.

A simple index was developed in order to objectively characterize a heat wave pattern in either analyses or forecasts of 500 mb height. 500 mb height was chosen because this parameter is usually among the most accurate fields predicted by numerical models. In addition, it is highly correlated with the surface air temperature on many time scales. The heat index IN is designed to be applied to any region of interest and to be used for any objectively analyzed data and specifically those of NWP models.

During the summer of 1980, a prolonged heat wave and drought affected the United States. A preliminary set of experiments has been conducted to study the effect of varying boundary conditions on the GLA model simulation of the heat wave. Five 10-day numerical integrations with three different specifications of boundary conditions were carried out: a control experiment which utilized climatological boundary conditions, an SST experiment which utilized summer 1980 sea-surface temperatures in the North Pacific, but climatological values elsewhere, and a Soil Moisture experiment which utilized the values of Mintz-Serafini for the summer, 1980. The starting dates for the five forecasts were 11 June, 7 July, 21 July, 22 August, and 6 September of 1980. These dates were specifically chosen as days when a heat wave was already established in order to investigate the effect of soil moistures or North Pacific sea-surface temperatures on the model's ability to maintain the heat wave pattern. The experiments were evaluated in terms of the heat wave index for the South Plains, North Plains, Great Plains and the entire U.S. In addition a subjective comparison of map patterns has been performed.

A series of simulation experiments is being conducted as a cooperative effort between the European Center for Medium Range Weather Forecasts (ECMWF), the National Meteorological Center (NMC) and the Goddard Laboratory for Atmospheres (GLA), to provide a quantitative assessment of the potential impact of proposed observing systems on large scale numerical weather prediction. For these studies an advanced analysis/forecast simulation system was developed which provides for a more realistic assessment of the impact of proposed observing systems than was possible in earlier studies. This system consists of four elements: (1) An atmospheric model integration to provide a complete record of the true state of the atmosphere (called nature); (2) a conventional data assimilation cycle that is used as the control experiment; (3) a satellite data assimilation that differs from the control in also including fabricated satellite data incorporated in the forecast-analysis cycle; and (4) forecasts produced from both control and satellite initial conditions. Comparison of these forecasts with nature provides an assessment of the impact of satellite data.

In addition to the SEASAT forecast impact studies previously conducted with the GLA model, a study was also made of the global analyses of surface wind stress, wind stress curl, and surface sensible and latent heat fluxes, which are generated as a by-product of the assimilation process. Both instantaneous analyses of these quantities and 15-day averages of each of these fields were found to be in good qualitative agreement with published fields.

The purpose of this paper is to quantify the impact of the FGGE SOS on the GLAS analyses and forecasts in the Southern Hemisphere, and to evaluate the relative importance of its components, and in particular, the impact of the loss of surface buoys. This discussion is limited to the forecast impact of the FGGE system upon a region between 28 deg S and 56 deg S, 50 deg W-80 W which includes essentially the southern cone of South America. Although forecast skill is verified against the NMC analysis, which used only VTPR but no TIROS-N data in the Southern Hemisphere, enough conventional data exists over this region to avoid biased verifications.

For this project, a version of the GLAS Analysis/Forecast System was developed that includes an objective dealiasing scheme as an integral part of the analysis cycle. With this system the (100 sq km) binned SASS wind data generated by S. Peteherych of AER, Canada corresponding of the period 0000 GMT 7 September 1978 to 1200 GMT 13 September 1978 was objectively dealiased. The dealiased wind fields have been requested and received by JPL, NMC and the British Meteorological Office. The first 3.5 days of objectively dealiased fields were subjectively enhanced on the McIDAS system. Approximately 20% of the wind directions were modified, and of these, about 70% were changed by less than 90 deg. Two SASS forecast impact studies, were performed using the dealiased fields, with the GLAS and the NEPRF (Navy Environmental Prediction Research Facility) analysis/forecast systems.

A study has been made of the accuracy of the GLAS fourth order forecast model's precipitation forecasts over North America during the first Special Observing Period (SOP-1) of FGGE and of the impact of the FGGE special observing systems on these forecasts. Fourteen 120 h predictions were generated using the coarse 4 deg latitude by 5 deg longitude version of the GLAS model from both the FGGE and NOSAT assimilation cycles. These forecasts were then verified against a detailed set of precipitation observations. Separate verifications were performed for precipitation accumulations or = .01 in., or = .1 in., and or = 1 in. occuring in 6, 12 and 24 time periods. In addition, three different methods of verification were applied. In the first method, model precipitation forecasts are verified against the greatest precipitation amount observed within a 4 deg latitude by 5 deg longitude gridbox centered on each gridpoint. For the prediction at a gridpoint to be verified only one observation within the gridbox is required. In the second method, a weighted mean of the closest observations to the gridpoint is used, provided that at least two observations on opposite sides of a gridpoint are available. For the third method, the model precipitation forecasts were interpolated to the observation locations.

A series of realistic simulation studies is being conducted as a cooperative effort between the European Centre for Medium Range Weather Forecasts (ECMWF), the National Meteorological Center (NMC), and the Goddard Laboratory for Atmospheric Sciences (GLAS) to provide a quantitative assessment of the potential impact of proposed observation systems on large scale numerical weather prediction. A special objective of this project is to avoid the unrealistic character of earlier simulation studies.

The GLAS physical inversion method for analysis of HIRS2/MSU data was implemented on the McIDAS. The method of analysis is identical to that used in processing global retrievals for six months of 1979, with the exception that the McIDAS retrievals are done on a higher spatial resolution, with one sounding attempted in every 4 x 4 array of HIRS2 spots, corresponding to roughly 80 x 80 km at nadir. This 4 x 4 array is further subdivided in 4 2 x 2 arrays with the sounding being performed utilizing all the spots in the single warmest 2 x 2 quadrant as sensed by the 11 micrometer window channel. Whether accurate retrievals can be performed on the high resolution grid without editing was determined. Different retrievals for a synoptic situation was compared and the guess dependence of the high resolution GLAS retrievals were examined. The orbit crossing the central United States at 1/21/79 09257 was studied. Radiosondes at 1200Z were used for comparison.