Dear colleagues,
We are investigating the voltinism of a species of dragonfly in two rivers in Catalonia (northeast of Spain). To study the possible temperature differences between the two streams we have installed temperature meters (hobos) that collect values every 15 minutes. How could we statistically analyze whether the differences observed between the two locations are significant? Is there any quick and easy method to do it?
Thank you very much for your attention and cordial greetings.
Ricardo Martín
I suppose the answer depends on your specific research question, but you could quantify temperature in each location using a metric like cumulative degree-days (based on daily mean temperature), using a base value that is biologically meaningful to your study organism. I've used cumulative degree-days as a predictor for growth of fish in mixed effects models. Otherwise, I would think simple statistical methods for comparing daily mean temperatures would work (e.g., t-test). I hope this is helpful. Article Life and death in a dynamic environment: Invasive trout, flo...
Ricardo Martin,
You may want to take a look at the article below, especially statistical analysis. It might be useful.
Article A Note on the Flowering of Ajuga orientalis L. in Relation t...
Regards
Dear Brian and Aristidis,
Thank you very much for your responses and for the articles you suggest. I'm sure they will be very helpful to us.
Best wishes,
Ricardo
In addition to the cumulative degree days and the temperature means mentioned above by Aristidis and Ricardo, may also want to look at the temperature extremes (i.e. daily high and low T's of diurnal temperature oscillations, and perhaps how daily means change with the seasons; i.e. seasonal lows and highs of mean daily T). Perhaps there are temperature thresholds (low and/or high) that can significantly affect dragonflies in your study areas?
Dear James,
Thanks for your suggestions.
We have detected that a species of dragonfly presents two types of voltinism in two rivers separated by about only 8 km. We are checking whether water temperature can have an influence on these differences. However, we do not know what the minimum temperature is that allows larval growth. The bibliography consulted indicates that, in general, odonates have thresholds between 8 and 12 ºC, so we take these values as a reference.
Your suggestions on the study of seasonal variations could be very interesting. We will perform these calculations as we obtain the corresponding temperatures.
Thanks again,
Ricardo
Assuming that the size of your data is more than 30, check with student t-test.
Dear Habtamu.
Thanks for your comments.
In the case of t-test, is it necessary for the data to follow a normal distribution and for the variances to be equal?
A t test is probably not the best test for comparing large sets of continuous, not discrete data. I suggest you consider using a paired t test, or a test more common for large continuous data sets that don’t depend on normality, like Nash-Sutcliff or similar comparison tools. The suggestions about comparing degree days and thresholds are excellent, but developing statistics for such comparisons will require some creativity. There may also be some aggravatig factors to consider like differences in water quality (degree of hypoxia, nutrient loading, sediment loading, pH, etc) should should be given some consideration. For example, the sediment concentration may be affecting the stream temperatures, or the nutrient concentration the stream ecosystem.
Dear William,
Thanks for your suggestions.
By applying the calculation of the NSE index to our data, we have obtained a value of -0.12. Given that the value obtained is less than 0.6, is it correct to think that the temperatures are clearly different in the two locations?
Thank you so much.
- Badges
- Science topic