Trend correlation with climate

[djhocking]

How do we detect the trend in abundance if it's driven by trends in the climate covariates and we have the climate covariates in the model to describe the interannual population fluctuations? There must be something about separating signal and noise in the time series world or dealing with correlated time series. Climate people must have this figured out. Going to talk to someone at the Northeast Climate Science Center.

[djhocking]

Ambarish Karmalkar at the Northeast Climate Science Center and I had a good chat. It seems like the best solution for our problem is to put a linear trend in the model and then detrend the climate data rather than just normal z-score standardization. The basic ideal will be to do a linear regression on the climate data then use the deviations from that model as the covariates in the abundance model. I'll close this issue once I get it worked out.

[djhocking]

@ThanHitt - do you have the unstandardized climate covariates that match Data_SeasonalClimateStd.RData or the means and sd used to standardize them (those would be needed for prediction anyway? I need the unstandardized data to detrend the climate data for the abundance trend analysis.

[djhocking]

@ThanHitt It looks like I have them in the BKT_COVS_ENV.csv file. They are missing 2015, but those seemed off even in the standardized form and we don't have any 2015 fish data in this anyway.