Sampling, Coding, and Storing Flood Plain Data

Report Summary: This report attempts to develop practical methods of sampling, coding, and storing data relating to the agricultural occupancy of flood plains in the United States, given the problem that many objectives must be satisfied. It appears that the most efficient sample for analysis of areal distributions, for making estimates of the areal coverage of phenomena, and for comparative analysis is a systematic, stratified, and unaligned point sample. The systematic feature of such a sample facilitates an unconventional use of punched card systems that has the invaluable property of maintaining geographic ordering of the data. Once prepared, the punched cards provide permanent storage in a compact format. A nominal scale is used for coding. The card system with nominal coding and geographic ordering facilitates comparative analysis of land-use, soil, slope, land-capability, flood-hazard, or other distributions by allowing easy preparation of contingency tables and rapid location of residuals in their geographic setting. Similarly, studies of changes through time are facilitated. Since the systems and operations described are simple and readily applied, it appears that they could be undertaken directly in the field. This has the advantage of suggesting additional questions and supplemental observations while field workers are out in the study area, instead of later in the office.