Computational considerations for collecting and using data in the equidistant cylindrical map projection and the bounds of sampling geographic data at progressively higher resolution

Abstract

The Equidistant Cylindrical Map projection is popular with digital modelers and others for storing and processing worldwide data sets because of the simple association of latitude and longitude to cell values or pixels in the resulting grid. This projection does not accurately display area, and the diminished geographic area represented by cells at high latitudes is not often carefully considered. A simple mathematical analysis quantifies the discrepancy in area sampled by cells at different latitudes. The presence of this discrepancy indicates that the use of this projection can induce bias in data sets when both sampling and reporting data. It is demonstrated that as the resolution requirements of input data for models increase, the necessity of providing data to accurately describe smaller cells, particularly at high latitude, will be a challenge.