In my specific case, I have worked with a data set from 2001 to 2018 which had to be split up in 2012 due to major updates in the data. Therefore, I am now dealing with a data set of 12 years (2001 to 2012) and a data set of 7 years (2012 to 2018). I have to predict how the climate in the given area is changing in the future (I have 17 models prepared).

I used a pattern scaling approach to superpose changes in daily climate variables to North American Regional Re-analysis (NARR) aka. my data points. That means I would only use 7 years in the future to predict the climate changes for the latter part of my data set. Hence, not really climate as we normally would discard 7 years of weather data as being nothing near climate data.

Now, the question is then, how would you compare future climate with the current climate on a temporal scale of only 12 years, let alone 7 years? Can you somehow account for this?

EDIT I am looking into wildfire ignitions in western America based on so-called fire weather indices (Canadian fire weather indices (FWI)). I have previously been able to predict these with logistic regression for 2001-2012 and 2012-2018. Now, I would like to see, how this may be in future climate.

Therefore, I gathered future predictions of FWI from 17 models at a 2.5-degree resolution. I then found the daily mean difference between the past (2001-2018) and the future (e.g. 2082-2099). I superposed these changes to the NARR, 0.5-degree resolution and assigned them to my ignitions. E.g. if the daily mean diff in temperature is 5 Kelvin for one pixel at day 150, and two ignitions occurred within this pixel on that given day, I would add the 5 K to the original temperature measured. Thereby, I have the structure of the NARR but with the mean difference of 17 models.

This would be fairly easy if I would look into the 18 years as a whole. But, due to the major revision in 2012, I have now two datasets split as mentioned before.

  • $\begingroup$ Hi Thomas, welcome to ES. Not sure if I understood correctly, definitely 7 years are too little records to make any speculation about climate. Could you provide more details in the kind of variables you are looking at, domain and spatial resolution? How do your data fit with reanalysis datasets? $\endgroup$
    – Nemesi
    Nov 10, 2020 at 11:39
  • 1
    $\begingroup$ Hi @Nemesi, thanks for the response. Does the edit help? $\endgroup$
    – Thomas
    Nov 10, 2020 at 13:04
  • $\begingroup$ Still, it is not clear to me how the two subsets of the 18 years differ. I assume some of the variables that you used to compute the FWI (humidity, temperature, precipitation, wind speed, right?) have been recorded (measured? assessed? estimated?) differently? Can you compare your data distribution with the distribution of the reanalysis to see if they mach somehow (or whether there is a systemic bias between the two distributions that you can correct)? (btw, NARR stops at 2014, maybe you could choose another one?) What about the 17 models? are you talking about GCMs? $\endgroup$
    – Nemesi
    Nov 10, 2020 at 15:13


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.