Create df with the mortality raster using moving window

[cliu229]

Hi Andrew and SAMC community,

Thank you for developing this great package!

I am getting some errors reported in the Model mortality data section. I replaced the road raster in the SAMC python example script with a nighttime light raster, but during the run I get the error "cannot allocate vector of size 4320.9 Gb", but my raster file is only 245.14 KB, how can I fix the error during the moving window?

The code and warnings for the runtime error are as follows:

b_mort <- data.frame(
roads = raster::extract(
raster::focal(nightlight_zz,
w = focalWeight(nightlight_zz, 990, 'rectangle'),
fun = "sum", na.rm = T), b_points),
mort = dplyr::select(b_points@data, mort = LIFESTATUS))

Error in h(simpleError(msg, call)) :
cannot allocate vector of size 1687.2 Gb

Thankyou for reading this！

[andrewmarx]

My best guess is that the 990 in focalWeight(nightlight_zz, 990, 'rectangle') is way too big based on the unit size for the CRS of your raster. If you plot your raster and look at the axes, the 990 in your function is 990 times the distance of 1 unit on your axes. So first make sure your axes make sense, then pick a distance based on that.

[cliu229]

Thanks Andrew I am appreciated your quick reply， if I adjust the value to be smaller, for example, 10, the error message will become:
Error in stop_subscript():
! Can't subset columns that don't exist.
x Column LIFESTATUS doesn't exist.

so I think it might be a problem with my raster data?

thankyou very much.

[cliu229]

When I run the label projection section, is there an error somewhere?

proj4string <- "+proj=wgs +lat_0=34 +lon_0=114 +lat_1=34.1 +lat_2=34.6 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs"
raster::crs(zz_land) <- proj4string

I get the following:
Error in sp::CRS(value) : NA

[andrewmarx]

This is an issue with dplyr::select(b_points@data, mort = LIFESTATUS). It's telling you that b_points@data (which I assume is a data frame), does not have a column called LIFESTATUS.

[cliu229]

Thanks for the guidance, I agree that this is really the problem.