You have navigated to David Blodgett's Web Resume. Thank you for stopping by to check it out.
Blog6/24/09 - It Begins.
I plan on using this space for keeping up to date. There may be an archive, but at the least some recent happenings and a status report will be here.
On with it then.
I am typing at you with some major injuries caused by an unfortunate bike accident. Over 3500 miles in the saddle with minimal trouble and boom, the 30mph over the bars flier happens. I guess I'm paying my dues for all the good luck I had. They give me an ETA to return to light cycling of about 6 weeks. First or second week of August, there goes the summer.
This all gives me a good chance to focus on some things that I might otherwise not gotten as deep into. My graduate studies in water resources engineering will certainly be furthered because of this injury. For better or worse.
In that realm, there is exciting work going on.
I have developed a calibrated and reasonably validated rainfall-runoff model of North Fish Creek. 100 distributed sub-basins are represented with fully distributed rainfall. The rainfall component is created from the NEXRAD Level III N1P and a rain gauge at the outlet of the watershed. A single point volume correction is applied to the entire radar field so the pixel over the rain gage matches the volume observed by the rain gauge.
The current work is focusing on generating an entire summer of distributed rainfall to be applied to a continuous simulation. This continuous simulation works reasonably well using a constant rainfall over the watershed, but some very small events are not represented while other small events have strange timing and total volume. The hope is that the fully distributed rainfall estimate will improve the small summer storm events.
A downfall of using the radar rain estimates is the low resolution of the numerical value of the radar pixels. Values are represented to the nearest 0.1 inch of rainfall. The reason for this is that rainfall amounts are reported as hourly cumulative estimates on a five minute refresh rate. A script has been written to extract five minute incremental rainfall from the hourly cumulative estimates. This technique works well for high intensity events, but is not suited to low intensity rainfall. Radar estimates of drizzle are often 0 but a rain gauge may pick up a quarter inch in a half hour. The effect the quarter inch has on the continuous soil moisture condition is tremendous though.
A script is in development that will take this dilemma into account. It will attempt to apply the valuable high intensity radar indicated rainfall distribution while applying the accurate low intensity regional rainfalls to the entire watershed.
More to come about this project sooner than later. Till next time, I'll be sitting in front of an air conditioner working on some crazy radar stuff.
