What DAYMET data should we use (temp/pcip) in the Mid-Atlantic model?

[jclbrooks]

Over what time period should data be collected/averaged? Should there be averages/maximums/minimums?

[jclbrooks]

Covariates on occupancy:

• Maximum annual temperature averaged over 30 years - represents critical thermal max

• Average annual temperature averaged over 30 years

• No PCIP data for this because it doesn't really represent flow; we would need an accurate drainage area to calculate this, which is not precise with the featureid's that are matched with each site

Covariate on detection:

• No temp data for this, water temperature probably affects them the most but it doesn't always correlate with air temperature

• Total amount of PCIP for 5 days before each sampling session - moving average

[djhocking]

############## STILL NEED TO PUT IN DAILY COVARIATES INTO THE 3D ARRAYS BELOW - will only work if add western maryland separately or have pass in long format with dates to join in with daily covariates then spread

---- potential easy strategy would be to get for non-maryland as a vector with a left join by featureid and date, the replicate that 4 times into a matrix that's the same for each pass. Then create maryland separately and bind_rows. I have all the daily covariates in the tempData object (should rename - used old code)