Procedure for determining % radiation blocked by canopy only -------------------------------------------------------------------------------- The following steps were performed after running the initial TOPORAD and ELEVRAD scripts to obtain insolation coverages. The crux of the issue has been to determine the amount of radiation blocked by canopy only, in order to make the necessary accurate temperature corrections. This involves several careful steps and is not as straightforward as simply running HemiView on the fisheye photos because those photos do not differentiate between canopy shading and topographic shading. 1. Compute % radiation blocked by canopy and topography using HemiView. Several parameters can be changed within HemiView to make it's output more 'realistic', and much testing was done in order to ascertain the best way to make its results as 'real' as possible. Actual radiation values could never be perfectly matched between IPW and HemiView (their methods of determining insolation are obviously different). Therefore, we decided to use HemiView only for % of radiation blocked, then use the (more reliable) IPW values for actual values. However, this was not as straightforward as it may seem, because we had to take into account the ratio of diffuse to direct (cloudiness) as it varies from month to month. We used the values obtained from the IPW scripts using the Bristow & Campbell equation as input into HemiView for '% diffuse'. The percents of total radiation determined to be diffuse were: JAN: .86 APR: .59 JUL: .67 OCT: .63 FEB: .77 MAY: .56 AUG: .37 NOV: .82 MAR: .67 JUN: .47 SEP: .42 DEC: .85 For exoatmospheric radiation, the value 1370 W.m^-2 was used (standard HemiView value, though different from IPW value; however, varying this made no difference in the percentage results). We used a constant transmissivity value throughout the year of .77 (it varies throughout the year in reality, but running HemiView with varying numbers did not change the percentage results noticeably; default HemiView value is .76, and .77 is the average transmissivity throughout the year as determined by IPW scripts). The Uniform Overcast Sky (UOC) solar model was used with HemiView, which assumes that equal amounts of diffuse radiation are coming from all parts of the sky. We used this model over the Standard Overcast Sky (SOC) model (which calculates diffuse radiation based on where the sun is in the sky) because the actual radiation values were much close to IPW's when using it. Again, using either produced very nearly the same percentage results. Finally, we treated every site as if it were at sea level in order for the transmissivity values to remain true; while the actual values obtained were thus far from accurate, this made no difference at all in the values of percentages. It must be mentioned that perhaps the fine-tuning of all of these parameters was ultimately an excercise in 'splitting hairs'. The variations in percentage values when testing various combinations were all well within the probable error encountered in the actual hemispherical photos themselves. Still, it is best to set things up as accurately as possible. The resulting percentages, then, take into account not only canopy blocking the direct rays of the sun throughout the year, but also the remaining portions of the sky blocked which affect the amount of diffuse radiation upon the site. 2. Compute % radiation blocked by topography only. The first value we computed was the amount of total radiation upon each site's pixel, taking topography into account. Next, we determined the total radiation upon the sites' pixels with no topography. Dividing the former by the latter, and subtracting this value from one gives the % radiation blocked by topography only: radiation with topography 1 - ------------------------------- = % blocked by topography only radiation without topography Again, this was not as straightforward as it may seem. First, it was crucial to treat each site pixel as a flat, horizontal surface during these calculations. Second, diffuse and direct proportions of insolation had to be taken into account for each month. Much care was taken to determine the precise pixel on the 50m DEM corresponding the actual location of each site. A special version of the TOPORAD script was used to determine the radiation with topography. This was exactly the same procedure as the original toporad.csh script (i.e., taking into account topographic shading and diffuse/direct proportions) except that a horizontal surface was used. The only reasonable way to obtain flat surface radiation with topography was to actually edit the DEM's pixel values such that the site's pixel was flat with no slope or aspect value. This was done by editing every immediate surrounding pixel to be the same elevation as the site pixel (this would make very little difference, if any, to the amount of topographic shading affecting that pixel). This new DEM was then used to make new gradient, sky view, and terrain configuration coverages. Determining radiation without topography was easier. A special version of the ELEVRAD script was run on completely flat DEM's corresponding to the site pixel's elevation (diffuse to direct proportions taken into account). As expected, the resulting values varied across the HJ Andrews, with the lowest sites having slighly greater radiation values than the higher sites. 3. Compute % blocked by canopy only. The final step was simple and consisted of simply subtracting the results of step 2 from step 1: % blocked by canopy/topo - % blocked by topography = % blocked by canopy only canopy_block_procedure.txt 11/01