CDO

Hi Cat,

what do you want to do? Do you want to interpolate the variable on the fine grid to a coarser grid? What do you mean with overlay?

-Karin

Hi Karin,
Thanks for the response.

I want to avoid interpolating the data if possible. Here is what I want to to (with a little bit of basic r code in it).

I have a high resolution 'baseline' climatology grid which is of size 4320 x 8640 x 12 (lat x lon x time(months)) (baseline.Tmax below).

I also have produced a lower resolution 'change surface' for a future climate scenario by subtracting the modelled baseline climatology (Tmax.h) from the future climatology (Tmax.f). This results in a climatology grid of size 180 x 288 x 12 (lat x lon x time) (Delta.Tmax below).

#baseline.Tmax = Climatology of observed Tmax from 1980-2009
#Tmax.h = Climatology of modelled Tmax from 1980-2009
#Tmax.f = Climatology of modelled Tmax from 2065-2095
#create the change surface
Delta.Tmax <- gridArithmetics(Tmax.f, Tmax.h, operator = "-")
#Then add it to the baseline
Tmax_modelled <- gridArithmetics(baseline.Tmax, Delta.Tmax, operator = "+")

What I want to do is add this change surface to the baseline climatology to get a future climate change scenario for this variable (Tmax_modelled above), but in a very specific way:

I have already tried 'coarsening' the baseline to match the resolution of the change surface, which results in a simple grid arithmetic to add the two together (which is what I did above. The coarsening was carried out using a cdo regrid command).

However, what I really want to do is to add the change surface to the baseline WITHOUT altering the baseline resolution. I think in order to do this, I would need a fishnet of identical size to the baseline, which would then be 'overlaid' on the change surface and the values in the change surface would be 'taken on' by the new fishnet. So, where there was only one grid cell of value 1.2 in the change surface, there may be (for example) 100 grid cells from the fishnet occupying that one grid cell, which would all assume the value 1.2.

Does that explanation make sense?
Is there a way to create this new surface with cdo?

Sorry, IMO this is not possible with CDO without interpolation to the same grid resolution.

Ok. Thanks anyway.