Are VolumeAreas wrong? Sarah Luongo & other examples
SimonDedman opened this issue · 7 comments
Try running code on various datasets and checking outputs & comparing.
See Sarah's email.
Issue is in scaleraster?
looping plot.R on scaled & unscaled individual asc's = exactly the same outputs. Might be expected.
Sarah Luongo movegroup run. R notes:
Sarah text notes: "after going back and reading the description for the move package getvolume function, I don't think you have to multiply by the raster size I think it does it for you in that function. I think they mean that that's what the function does and not what you have to do"
"these [getvolumeUDareas] are way too big for snapper I think:
[movegroup results say] core is 0.213km2 and general is 1.02km2
they spend almost all of their time on the same reef
the move package (without multiplying raster size) was core area 0.0004km2 and general 0.0017km2
like 400m2 feels spot on for their core area
they just sit on the ARB artificial reef all day
and then at night they move probably 500-1000m to a reef nearby"
do we know that rasterres == rasterResolution? L529
rasterres 2500, rasterResolution 50. 50*50=2500
rasterres = rasterResolution^2 so larger rasterResolution will be more disparate between move() & movegroup()?
L473 could add resolution = rasterResolution
L517 xAEQD should have resolution = rasterResolution
L528, confirmed, rasterres is resolution(xAEQD)^2 which means resolution(xAEQD)=50
L827, core/general SHOULD be multiplied by rasterres i.e. rasterResolution^2? rasterResolution^2 = the area of 1 cell. So Core/general should be the COUNT of cells?
I tested rasterResolution in a loop and it wasn't the culprit
Data we have:
- Sarah Luongo
- STB hammerheads [SD/Vital/TG to check]
- Rob Bullock lemons [Rob to check]
- Vital's hammerheads bimini [Vital to check]
- Mo 8 species [table data not used except Tigers checked and ok] [Mo to check other 7]
- Liberty Boyd turtles [Lib to check]
- Blocklab tuna [SD]
7 (14) datasets.
SD to email Mo then everyone asking them to check
movegroup.R L748 block: switch from getvolumeUD() & back.
scaleraster.R L254,8, 284,5: still using max not getvolume.
Could compare outputs from scaleraster under these 2 conditions, with these 2 version of the code approach
Vital hammerheads: tested 2 sharks, values are correct.
No bug. Calcs are fine.