search

drnickisaac / sMon-Amphibia

Analysis of trends in German amphibia
3 stars 2 forks source link

Numbers of records in occDetMod output do not match sum of records in Regions #4

Open EichenbergBEF opened 6 years ago

EichenbergBEF commented 6 years ago

@drnickisaac Hi there, I was just having a look at the results of the amphibia models with the federal states included as a random factor. I found the following thing: the model output for the species (e.g. Alytes abtetricans) give a nuRecords of 317 total records.

However, when I have a look at the reported number of occurrences in the federal states, separately (by summaryzing the raw data to the reported occurrences of species according to federal states), I derive a total of 290 occurrences.

This discrepancy is found in all species (sometimes with a difference of more than 50% between the model outputs and the calculated sum per region and species).

The most extreme is for Buf bufo where the numer of records in the model are given as 19873, whereas the single federal states sum up to only 5582 records.

I am wondering if I am missing something here. Can you help me to understand out where the differences come from?

For a better understanding of what I did, I attached a short Rscript which you can run using the model outputs from the data portal.

short_script_for_in_depth_exploration_amphibia.txt

bowlerbear commented 6 years ago

you did: length(unique(ENDE) Do you get the same if you remove unique? currently you are removing counts on the same day - i wonder whether this explains the difference.