I begin looking into temperature dependence of LEDs, and found that temperature does have an effect on wavelength. However, if I can identify the range of temperatures my sensor will be exposed to, it may not have a significant effect.
According to one paper by Chhajed et al. (2005), the peak wavelength for UV LEDs only changed by about 5 nm over a 100 degree C temperature range. The temperature range of my sensor will be far less then that. However, the sensor will likely be in temperatures below 20 degrees. I will look for someone who has investigated this relationship down to lower temperatures, but I’d expect the same relationship to continue beyond 20 degrees. Red LEDs appear more sensitive to temperature. Although, I am not sure if the same relationship will hold true for when LEDs are used as photodiodes.
I also looked into data loggers and found one that should work well. http://iteadstudio.com/produce/arduino-sd-card-shield-and-filelogger-library/
Here is another link to instructions for building an amplified color sensor: http://www.robotroom.com/ReversedLED.html
Chhajed, S., Xi, Y., Gessman, Y., Gessman, L., Schubert, E. F. (2005). Influence of
junction temperature on chromaticity and color-rendering properties of
trichromatic white-light sources based on light-emitting diodes. Journal of Applied Physics. 97: 054506.
