Run CABLE-POP at ACCESS Resolution

Land Surface CABLE
21

It seems like the directional winds are only used in the vector magnitude calculation, so it should be fine to just pass the sfcWind and a dataset of zeros and get the correct result. I’ll give that a go.

22

I’ve run the climate restart stage with the ACCESS met forcing and the results are quite promising. I’ve attached plots of the biomes computed using the ACCESS ESM1.5 data and the CRU-JRA met data as used in the TRENDY runs, and it looks pretty good.

ACCESS_Biomes
ACCESS_Biomes1920×1043 177 KB

TRENDY_Biomes
TRENDY_Biomes1920×1043 143 KB

The bits and pieces I had to do to achieve this with the ACCESS ESM1.5 data is as follows:

  • Resample the surface pressure data from a 6 hour frequency to daily using using the mean over the day
  • Convert the ACCESS1.5 precipitation from mm/s to mm/day
  • Create a dummy wind file filled with zeros to use with sfcWind
  • Reduce the resolution of the nitrogen deposition file from the CRU-JRA resolution to the ACCESS ESM1.5 resolution
  • Create a landmask for the ACCESS ESM1.5 resolution. This was somewhat involved, as there were issues with the ACCESS points landing on islands in the middle of the ocean that the 1 degree resolution gridinfo_CSIRO_1x1.nc skip over, so the nearest neighbour search failed as the nearest point in the existing gridinfo file fell outside the search radius. To avoid this, I removed all islands from the generated ACCESS landmask, where I defined an island as any land point surrounded by non-land points (including diagonal neighbours).
23

Interesting that ACCESS has more of the continent covered in trees - I’m wondering if that’s a rainfall bias in the model, though I thought that ESM1.5 was typically drier than most models, or at least dries more with warming than other models.

24

I’d accidentally used the iveg rather than the biome variable, I blame it on it being too late. I’ve edited my original message to have the correct biome plots.

25

I haven’t looked into the details as to how the biomes are computed, I would have to defer to @Juergen on that one.

26

The biomes in CABLE-POP are calculated in subroutine biome1_pft in cable_climate.F90. They are based on Prentice et al. J. Biogeog., 19, 117-134, 1992

27

The full physics spin-up appeared to be successful, I was just making an embarrassingly stupid error (had switched StartYear and EndYear). I would like some suggestions as to what variables should be interesting to look at, in any of the output files (cable, CASA and LUC). Unfortunately the CASA and LUC outputs don’t contain any descriptions, so difficult for me to interpret.