Summer 2004, I was lucky to get a Summer Research Scholarship to ANU in Canberra. It was my first time in a proper research lab. The train arrived in Canberra and I was picked up by their lab tech who was a real motherly type of lady who had been with the group for 10 years. My lasting memory of her was her telling me that research never goes as quickly as you think it should. Words to live by! My project is a little like that at the moment. It’s not going as quickly as I had hoped, and annoyingly, my frustration is affecting this too – not picking up on simply errors in my code.
Today I edged forwards a little, finishing a Python script to calculate WBGT. I have re-compiled Matthias’s C# code and checked the two against each other for consistency and all is looking good. There is nothing wrong with Matthias’s C# code and ultimately, a compiled language is going to run faster than an interpreted one. But the simplicity and my familiarity with Python means I can work more quickly. I am already familiar with Python’s MySQL connector and graphing tools, which will a big help in the weeks to come.
The algorithm for Tw (wet bulb temperature)
The algorithm approximates Tw by iteratively refining the value using the dewpoint temperature, Td, as the initial estimate. The Climate Chip group are using Liljegren’s WBGT model, though to be honest I have been struggling to figure out how they have gone from Liljegren’s model to the formula in their code. I have tried substituting in the appropriate constants and simplifying Liljegren’s equations to match Matthias’s, but it defeats me! So I thought to code up Liljegren’s formula and run it against Matthias’s to visually see that they were equivalent, but I need to track down a couple of additional constants relating to the angle of the sun to do this. So for now I will take it on faith, and ask Todd if he can introduce me to Bruno Lemke so I can discuss it with him.
Steps from here
Step 1: Code up the UTCI calculation
Step 2: Extract the experimental measurements (longitude, latitude, year, month and temperatures) to file
Step 3: Connect the Python script to a MySQL instance
Step 4: Calculate WBGT and UTCI, loading the resulting rows into the db.
Further: begin to look for patterns in the UTCI and WBGT calculations, see if it is possible to prematurely short-circuit these and still produce an accurate result.