Earth Space Research
Group Earth Space Research
Group
One recent report of ship based measurements indicate that a severe ozone event can reduce rates of phytoplankton primary production by up to twelve percent (Smith et al. 1992). Such in situ measurements are valuable for pinning down the mechanisms which relate the average surface flux of UV-B radiation to local phytoplankton production rates.
To tie these specific findings to an understanding of how ozone depletion affects the entire Antarctic ecosystem one must somehow provide a global estimate of the UV flux. Measurements from the TOMS instrument on board the TIROS 7 satellite have been successful in determining the temporal and spatial distribution of ozone depletion . Under cloud-free conditions this information is sufficient to determine the surface flux of UV-B radiation (the solar flux of UV-B impinging on the top of the atmosphere is fairly constant and well known). However, the maritime Antarctic region is one of the most chronically overcast regions in the world and the cloud cover will reduce the strength of the effect we are trying to measure.
We can provide the required global estimate of UV-B surface flux by using a radiative transfer computer code to model how much UV-B radiation penetrates the atmosphere. The estimated radiative flux will depend not only on the cloud optical depth and ozone column density but also on whether we are considering an ocean or snow covered surface. Multiple reflections between a highly reflective snow surface and cloud layer can nearly cancel the flux attenuation due to cloud cover.
Figure 1 shows an visible image returned from the NOAA 11 meteorological satellite on 8 Oct, 1991. The field of view is centered on Anvers Island which is off the north west side of the Antarctic Penninsula. For each pixel in this scene the cloud optical depth can be evaluated by comparing the pixel brightness to a theoretical curve generated from the radiative transfer model. For a given surface reflectance, the brightness of a pixel goes up with cloud optical depth .
Unfortunately, this approach will not work over a highly reflective snow or ice covered surface. In this case thickening the cloud layer makes very little difference to the visible brightness observed from space. In this analysis we tried to infer a relationship between the cloud optical depth and the cloud top temperature, which could be determined from satellite infrared measurements. We hoped that this relationship would allow us to deduce the cloud optical depths over a snowy scene using the thermal measurments.
Figure 2 shows the surface reflectance for the same field of view as Figure 1. The image is a composite of cloud-free regions taken from several satellite passes in October of 1991. The information in Figures 1 and 2 are combined to yield a cloud optical depth value at each pixel as described above.
Finally, on a pixel by pixel basis, the cloud optical depth, surface reflectance and ozone column depth is fed into our RT code to obtain the results shown if Figure 3. In this figure notice that the UV-B flux is larger over the snow covered surfaces. This is partially due to the multiple reflection effect mentioned previously. But part of the contrast is due to the weakness of our assumed correlation between of cloud top temperature and cloud optical depth. In future work we will try to develop better methods to retrieve the cloud optical depth over snowy surfaces. Other areas of investigation will be 1) better ways to obtain surface reflectance, 2) corrections for non-plane-parallel clouds and 3) methods to handle non-lambertion surface reflectance.
Return to ESRG overview
