Visualizing Sipaliwini weather data

[joneill809]

I'm kind of a data geek (my day job) so I was interested in coming up with some data visualizations for weather information collected where our frogs (azureus) are from to see how accurately my viv climate matches up with their species' natural habitat. Now granted these are CB frogs that are probably decade(s) removed from the wild, but it was still a point of curiosity for me. I'm sure this data is common knowledge, but I figured you would enjoy seeing the visualizations from this weather station.

Some Googling quickly located a weather station in the SIpaliwini Savannah Nature Preserve with data of varying degrees of completeness. I pulled all the data, and summarized the temperature and humidity statistics by month for all available data points from 1973 through 2011. The box plots from these summaries are:
[ATTACHMENT NOT FOUND]

I prepped a more comprehensive view of the data that includes a summary year to year of what data points were available. It's kind of big, and I made it into a JPEG so I could upload it. I have the PDF and a CSV of the actual data if anyone is interested. This file contains a view of the weather station's location within the preserve. I've never been to Suriname so I'd like to know if this weather station looks like it's in a representative location. The zoomed in view of the coordinates seem like they are in a forested area.
[ATTACHMENT NOT FOUND]

Some observations:

• Median monthly humidity at this site varies from 70% to 83% with lower humidity values in the hotter months.
  
• Average temperatures by month are very stable and hover in the high 70's. Average lows around 70 and average highs in the high 80's to lower 90's.
  
• Data collection reliability at this site is not so good. There was nearly a decade gap in data collected from 87 through 96, and most years are quite spotty - yep that big white space in the first time series plot on the upper left of the large visualization is a data gap.

I think most of these observations correlate quite well to the rules of thumb I have seen on DD - specifically "mid 70's" for the target temperature and 80% humidity.

Please provide feedback! I'm looking to iterate on this and make improvements. I was mostly interested in temperature and humidity, though limited rainfall data is available. Due to how incomplete the data is I'd probably have to narrow the focus of rainfall analysis (cumulative for the month) to the mid 80's where the data was most complete. If enough folks find this useful I can write up some macros to automate the compilation of weather data and plotting so we can visualize more stations in areas of specific interest.

[rcteem]

Awesome post!!!! Thanks for taking the time do share this with the hobby.

[joneill809]

Awesome post!!!! Thanks for taking the time do share this with the hobby.

Thanks Chris! I'm going to script this so I can run the same sort of analysis for other weather stations (probably a week or two of tinkering at night). The best I think we can do is to look up the coordinates of the available weather stations and grab something that is in close proximity to the locale of interest, though I don't know how close we will really get.

However, longer term, if we do something like profile all of Suriname, Peru, CR, where ever, we may be able to get a better representation of the overall regional patterns. That's really what's driving my automation interest - what would the visualization challenges be to expanding a summary view to each country or the continent? Fun stuff!