I need some gridded population information so I downloaded GRUMPv1, which seems like it should be perfect for my needs. However the number in the file naup00g.asc (North America unadjusted population counts in the grid) are small floating point numbers. I had expected that they would be population counts in the grid cell, which is what the documentation seems to say. Not sure how to interpret values like "2.35144e-005" as a population count! Am I misinterpreting these somehow?

1) GRUMP uses a proportional allocation algorithm to distribute population across a landscape based on land area that also takes into account whether a pixel is classified as urban or rural.

2) If an administrative unit has a large amount of land area, but a small population count, the result might be that certain pixels are coded with floating point values of less than 1 for population count. This is an artifact of the proportional allocation algorithm.

3) Due to this fact, the GRUMPv1 dataset (or any global dataset of a coarse resolution) is not appropriate for pixel level analysis. What is more appropriate is to examine the data in terms of locality (neighborhood). If you are doing a local study then it would be best to look for a higher spatial resolution local dataset.

Feedback and Knowledge Base