Introduction of ROI Selection for GOES Products Inter-Satellite Comparison

Purpose of Inter-Comparison Between GOES-East and GOES-West

Inter-satellite comparison of products from GOES satellites at different locations provides key information for quantifying the temporal consistencies of the products. It also allows evaluation of the attached precision over a larger range of situations compared with using ground measurements as a reference. The time-series comparison results may also provide an independent way to build a community reference, as benchmarking of L2 land products is essential to identify and possibly resolve differences between algorithms and upstream input data, leading to improvements in their accuracy and reliability.

For any potential major discrepancy beyond the seasonality of the products that fails to meet users’ requirements, the product team will investigate the reason according to the assumptions embedded in:

• Processing flow including radiometric calibration, geometrical correction, projection
• The retrieval algorithm
• The sensor characteristics (spectral function and point spread function)
• Snow and cloud detection

Region Used in the Cross-Comparison

A region bewteen 100°W and 90°W in longitude and between 30°N and 40°N in latitude was selected for this purpose based on the following criteria.

Region Selection Criteria

The cross-comparison is expected to be fair and objective and suffer less from uncertainties, such as:

• Less influence from the terrain effect
• Smaller View Zenith Angle (VZA) difference

The first factor is quantified by coordinate shifts due to the parallax effect caused by terrain relief (i.e., the effect of change in elevation). These errors are more obvious in the western United States, where elevation first rises eastward from the coastline over the coastal area and then the Sierra Nevada, as well as in the ice-capped Andes mountains in the east of Chile. Less terrain effect is expected over the area with a smaller distance. Figure 1 shows the map of the first factor calculated from geolocation coordinates before and after the terrain correction. The terrain-corrected geolocation data is courtesy of Chris Schmidt from the GOES-R fire product group.

The second factor is assessed through the VZA difference between GOES-east and GOES-west. To ensure geolocation matching, the VZA maps are re-projected and compared on a 0.02° equal Lat/Lon grid as well as the terrain-caused location shift distance. Figure 2 shows the VZA difference map.

The final indicator for suitability is the weighted average of the two factors normalized to be between 0 and 1. Figure 3 shows the map of this indicator, as the basis of the region selection used for this comparison.