Modeling Surface Photosynthetic Active Radiation in Taylor Valley, McMurdo Dry Valleys, Antarctica

• Main Author: Acosta, Dimitri R.
• Other Authors: Berkelhammer, Max, Dombard, Andrew, Doran, Peter
• Format: Thesis
• Language: English
• Published: 2016
• Subjects: ice-covered lakes; geographical information system (GIS); R model; meteorological data; digital elevation model (DEM); Bonney; Fryxell; Hoare; Lake Bonney; Lake Fryxell; Lake Hoare; McMurdo Dry Valleys; Taylor Valley; Antarc*; Antarctica
• Online Access: http://hdl.handle.net/10027/21180

Primary Productivity within ecosystems is a core research area of the National Science Foundation’s Long-Term Ecological Research Network. This study presents T-sPAR, a computer model that estimates total daily surface Photosynthetically Active Radiation (PAR) for Taylor Valley, located in the McMurdo Dry Valleys of Antarctica. T-sPAR incorporates 20 years of meteorological records from four stations to constrain a topographically corrected surface radiation model. A locally weighted regression model that included 84% of all daily maximum observations was used to determine expected total daily PAR under cloudless conditions for each station. A simple cloud cover model that examines the difference between total daily measured and expected PAR was used to bin atmospheric conditions into four categories: clear, scattered, broken and overcast. The cloud cover model shows a positive correlation between proximity to the ocean and seasonal overcast conditions. Meteorological data indicates a positive correlation between total annual surface PAR and valley width. T-sPAR estimates indicate that north facing slopes receive the highest amount of annual surface PAR while south facing slopes receive the least. Lake surface estimates suggest Lake Fryxell receives the most seasonal surface PAR and has the narrowest range in values across its surface. Lake Bonney has the widest range in annual values across its surface, and Lake Hoare receives the least total amount of surface PAR due to consistent topographic shading. Bi-monthly surface PAR maps created with T-sPAR show the distribution of surface PAR at time step intervals that can be used in the field to identify future sample sites. Finally, a user interface provides tailored estimates for any day in the record, correcting for atmospheric conditions based on user input.