Project

General

Profile

land cover changes in COSMO-CLM

Added by Bo Huang about 2 years ago

Dear all,

My research group wants to set up new simulations by changing existing land cover (e.g., deforestation experiment: forest→bare area). We therefore need to change the external file variables. Our approach is to take the parameters of a grid with “bare area” (fraction > 0.9) and use them to replace the parameters of the forested land class in the grids where forest is dominant. We identified the following parameters connected to land cover to be changed:

var2 = (/“PLCOV_MN",“PLCOV_MX",“LAI_MN",“LAI_MX”, “FOR_E”, \
“FOR_D”, “PLCOV_MN”, “PLCOV_MX”, \
“Z0",“ROOTDP",“NDVI_MAX",“NDVI",“LU_CLASS_FRACTION”/)

Is it correct? Does anyone have any suggestion or comment on it? Thank you very much.

Greetings,
Bo


Replies (4)

RE: land cover changes in COSMO-CLM - Added by Burkhardt Rockel about 2 years ago

My first guess for bare soil would be as follows (the soil experts may correct me):

PLCOV_MN(rlat, rlon) ; = 0
PLCOV_MX(rlat, rlon) ; = 0
LAI_MN(rlat, rlon) ; = 0
LAI_MX(rlat, rlon) ; = 0
Z0(rlat, rlon) ; z0_bare = z0 - z0_veg (z0_veg is only available in the ECOCLIMAP option in WebPEP, therefore run WEbPeP a second time choosing "ECOCLIMAP" for "Land Use", do a  Z012(1) - Z0_VEG(1) from that data set and replace the z0 in your external data set by the z0 you got)
EMIS_RAD(rlat, rlon) ;=  0.99
RSMIN(rlat, rlon) ; ?? 0 ??
FOR_D(rlat, rlon) ; = 0
FOR_E(rlat, rlon) ; = 0
ROOTDP(rlat, rlon) ;=  0
NDVI_MAX(rlat, rlon) ; =  0
LU_CLASS_FRACTION(nclass_lu, rlat, rlon) ; NA (is not used in CCLM presently)
ALB_DRY(rlat, rlon) ; ?? this is for itype_albedo=2. I suggest to use itype_albedo=1 in your case, then the albedo is dependent on the soil type
ALB_SAT(rlat, rlon) ; ?? this is for itype_albedo=2. I suggest to use itype_albedo=1 in your case, then the albedo is dependent on the soil type
NDVI(time, rlat, rlon) ; = 0
NDVI_MRAT(time, rlat, rlon) ; = 0

RE: land cover changes in COSMO-CLM - Added by Bo Huang about 2 years ago

Hi Burkhardt,
thank you very much for your suggestion.
Firstly, I followed the CORDEX-EU settings. The land cover dataset is from GLC2000. Because I found that the new version of COSMO has not joined the CORDEX-EU. Then, our findings can contribute to the CORDEX-EU project.
Then, I found that for bare land grid (e.g., Sahara desert),
PLCOV_MN = 0.02
PLCOV_MX = 0.05
LAI_MN = 0.4
LAI_MX = 0.6
Z0 ~= 0.05
EMIS_RAD = 0.95
RSMIN = 150
FOR_D = 0
FOR_E = 0
ROOTDP = 0.3
NDVI_MAX = 0.11~0.13
NDVI = 0.11~0.13
NDVI_MRAT = ~0.98

We try to replaced these parameters of a specific forested grid (if (maximum LU_CLASS_FRACTION ⇐9)) with that from a representative bare land grid. This is only one case. We also want to change forest to grass. The approach is the same. Changing the parameters of forested girds with that from a grass land grid. I am not sure this is a correct method or not. Maybe the soil experts have comment on it. Looking forward your reply.

Greetings,
Bo

RE: land cover changes in COSMO-CLM - Added by Hans-Juergen Panitz about 2 years ago

Hi Bo,

I suggest to contact Merja Toelle (University of Gießen, and of course CLM-Community member).
She has a lot of experience with this aforestation/deforestation stuff and the necessary manipulation of external data sets.

Hans-Juergen

RE: land cover changes in COSMO-CLM - Added by Bo Huang about 2 years ago

Hello Hans,

thank you for your information. I will contact with her.

Greetings,
BO

    (1-4/4)