markus_todt@northumbria_ac_uk
New Member
Hi allI'm currently looking into the sub-canopy longwave radiation beneath boreal forests and as part of this I'm calculating the vegetation temperature and sub-canopy longwave radiation as in CLM4.5 but for observational site(s). Basically, I copied the required parts of the CLM code and run it with site observations - an alpine site in Switzerland - in my own little toy model (which also includes another model's code for comparison). If there's no mistake in my toy model the vegetation temperature and consequently the sub-canopy longwave radiation reaching the ground, at least for needleleaf evergreen boreal trees, seem to have a very large diurnal cycle for "clear-sky conditions", i.e. when there's a large discrepancy between the incoming atmospheric longwave radiation and the actual air temperature above the vegetation. Compared with observations for the sub-canopy longwave radiation, this results in an overestimation during daytime and underestimation during nighttime - for daily means those errors cancel each other out to a certain degree.Right now, I'm trying to get PTCLM running so I can validate my toy model but I haven't been successful with that so far. So, at the moment the only way is to compare the results of my toy model with actual global simulations for CLM4.5. The attached plot displays the difference between the air temperature above the vegetation (TBOT for CLM4.5 and tower measurements for my toy model) and the vegetation temperature against the ratio [incoming atmospheric longwave radiation / theoretical longwave radiation for air temperature above the vegetation], so a measure of clear-sky (low ratio) and overcast (high ratio) conditions. For the global simulation I used hourly values for January-June of the 1972 year under Quian forcing, the site data are hourly as well for January-May. The toy model results are within the range of the actual CLM4.5 simulations, although at the end of the spectrum. For context, I also calculated a PDF of the global simulation values below an effective sky emissivity of 0.8 and you find that plot attached as well. There certainly are two peaks in the PDF, one for nighttime values with negative temperature differences and one for daytime values with positive temperature differences (and one large peak just slightly below 0). Has anyone encountered similar issues or looked into a similar subject? CheersMarkus