Posted on June 21, 2013 by James Douglas
Today the goal was to try to figure out what was wrong with the elevation data from yesterday and try to see if more accurate elevation data could be determined. We drove to Dr. Becker’s rented house right outside the ranch and had a discussion with him about what had gone wrong with the data from yesterday and how it was uncertain that even with post-processed data that the elevations garnered from the Trimble’s would yield results that would significantly improve the DEM. Because he didn’t have enough field experience with the GPS units, me and Julianna were referred to Dr. Wechsler, who after a fair amount of familiarization with the data, agreed that we were probably right but that we should correct the data before making any definite conclusions. However, while talking with Dr. Wechsler, she suggested I become more familiar with water budgets on my own and a cursory google search revealed an ArcToolbox for generating water budgets. This relieved a lot of pressure for me, because now I had a defined and tested methodology for pursuing my research question. So it was back to Dr. Becker’s (who had internet) to post-process the data. After the processing, the data still looked unusable. We worked with Paul for a while to see what data points were sensible and which ones were most likely erroneous, but there seemed to be no systematic way to eliminate data or any real pattern at all to the errors. This led us to the conclusion that the GPS unit itself must be faulty, so we decided to test this theory by retaking the most erroneous points with another unit. This consumed the rest of the field day, as at first the unit wouldn’t function properly, and after the issues were hammered out only an hour of field time remained.
Later that evening, we performed a differential correction to the new data, and the results were slightly more accurate (seeming to support our earlier theory about the faulty unit) but were nonetheless still unusable as a method for correcting the DEM. This left me and Julianna relatively frustrated, and the only reasonable conclusion from two (invaluable) days of field work was that the methods that we had pursued were fruitless. Luckily for me, a water budget could most likely be effectively estimated from existing data, but I sympathized with Julianna, whose research question functioned on a much smaller scale, and whose project could have benefited from more accurate elevation data. Dr. Wechsler spent some time with me delving into the NOAA datasets of precipitation data for Oahu, but we quickly became too exhausted to make any significant progress. Tomorrow my plan is to focus on conducting research into finding existing datasets and/or researching ways to make existing data work for my purposes in estimating a water budget for the Ka’a’awa valley before time runs out.