In Mississippi, data sources tended to conform in different areas for the two main classes. In the delta region, estimates for agricultural land from any one source were, on average, within 10% other sources, but estimates for forested land not within 60% of one another. The pattern was inverted in the rest of the state, where one estimate for agricultural land was within 10% of another for only 38% of the counties and nearly 80% of estimates for forested land were within 15% of another estimate. Agricultural and forested land estimates agreed most often for the 1992 NRI and MRLC datasets, with 27 and 49 counties (33 and 60%) within 15% and 30% of one another for agricultural land, respectively, and 66 and 87 (54 and 71%) for forest estimates. The MRLC and AVHRR agreed least often with only 20 counties within 15% of one another for agricultural land and one county with an MRLC estimate of forested land within 15% of the AVHRR estimate. Only 24% of counties had one data source within 15% of another for agricultural land in Nebraska, and nearly two-thirds of that agreement was between the 1992 NRI and AVHRR. Rangeland area estimates from one source were within 15% of those from another source for only eight counties. Estimates of forested land in Virginia were similar for different data sources, with 23% of estimates within 15% of another estimate and 76% within 30%. Most (95%) estimates from the 1992 NRI and the MRLC datasets were within 30% of one another. Coincidence between data sets was most common in the more mountainous MLRAs where, according to the 1992 NRI, forested land covered an average of 60% of the area.
