Determination of the proper site spacing density over Texas
The current meteorological surface observation network in Texas only monitors one out of every five counties. The observation sites are nominally spaced 150-200 km apart and report hourly measurements. For this reason, the current system is incapable of providing observations at a spatial and temporal resolution to document mesoscale weather features and provide short-range forecasting. To overcome this problem, various site spacing procedures are used to propose to fill the gaps in the monitoring system.
In this study, site spacing determination procedures based on correlation level, power spectrum, and true field error variance are used to analyze the temperature, dew point temperature, wind speed and pressure parameters over Texas. Hourly observations from 126 surface observing sites located in Texas and the adjacent states with a data coverage period varying from 6 to 21 years are considered for the study. The existence of isotropic conditions over the domain is tested by examining the spatial correlation variations of the parameters. The existence of anisotropic conditions for each parameter resulted in the search for sub-regions within the domain. Cluster analysis indicated three separate clusters for the domain. For each cluster, the spatial correlation variations are examined and spectral analysis is applied to determine the governing scales for each parameter’s variation. Error amounts in obtaining the true Fourier coefficients are analyzed; the highest wave number that can reasonably estimated is determined for each parameter in each cluster. Finally, the error variance in determining the true field for each parameter is examined for site spacings of 200, 150, 100, 50, and 25 km for each cluster.