I'm working on an analysis that involves summarising a stack of unprojected (WGS84) global 30" climate rasters (various expressions of temperature and precipitation), and generating final output in an equal-area projection.

Is it better to project the individual global rasters first (e.g. to Mollweide) and then summarise, or to summarise and then project? What factors (other than computational efficiency) should be considered in this regard?

Additionally, final output will probably be aggregated to 10 km. Is it more appropriate to aggregate raw data to, say, 0.1° and project to 10 km, or project to ~1 km and then aggregate to 10 km?

  • 2
    Summarize pre-projection. Projection is a resampling and will modify the values slightly so the data is less accurate. – Michael Stimson Mar 21 at 0:26
  • Thanks @MichaelStimson - regarding my last Q, would you consider aggregation part of the summary process (and do aggregate before projection)? Does order matter here? – jbaums Mar 22 at 13:48
  • 1
    Aggregation is also in the same family as resample, though if you're using Esri Aggregate resources.arcgis.com/en/help/main/10.2/index.html#//… the cell_factor option is an integer (long). It depends on what you're hoping to achieve, a good aggregation smooths out irregularities so it might help to aggregate to a mean or sum first then summarize to 10 times the cell size for your tabular statistics. One thing to note though is you don't really need to project your rasters to create your maps, every GIS package I have used in the last 10 years will project on the fly. – Michael Stimson Mar 24 at 23:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.