Arctic_50N pixbypix

Colorbars are in the same location (filename starting with 1_)

I calculated the following metrics:
climat = multi year (1998-2018) climatology, simple mean of valid pixels (at least one valid year needed)
climat_min4yr = same as climat, but at least 4 valid years needed.  Basically a subset of climat.  Trends are calculated for these pixels only (4 points needed to calculate p-value).  "Laptev Sea bloom" doesn't show up here (only 1 valid year).  See http://albedo.stanford.edu/gertvd/research/arctic/arctic_50N/pixbypix/prod_tot_num_valid_years_1-10_arctic_50N.png to see all pixels with less or equal to 10 years of valid data.  Color bar: http://albedo.stanford.edu/gertvd/research/arctic/arctic_50N/pixbypix/1_colorbars_pval_validyears.png
slope = slope of linear regression (trend) 1998-2018
slope_significant = pixels only plotted when fit is significant (p < 0.05)
p_value = p value of fit
num_valid_years = number of points (valid years) that were used for linear regression.

For these variables:
prod_day1 = first day of the year when pixel has a valid production value
prod_daylast = last day of the year when pixel has a valid production value
season_len = production season length:
prod_daylast - prod_day1 + 1
season_len_ow = production season length weighted by open water fraction: 
∑ (daily open water fraction between prod_day1 and prod_daylast)
prod_val_fr = fraction of valid production days during production season.  If value 1, than no missing production values between prod_day1 and prod_daylast.  If value low then lots of missing production values in the time series.
# of valid prod days / season_len
prod_tot = annual production per square meter (gC/m2/year).  Calculated as mean daily production x production season length (season_len).  Mean daily production is weighted by open water fraction: 
prod_mean = ∑ (daily prod x daily open water fraction) / ∑ (daily open fraction for days with valid production)
prod_tot = prod_mean x season_len
chl_mean = mean Chl concentration between April 1 and October 15 (198 days).  Days with missing Chl are ignored (ie. not counted as 0):
∑ (valid daily Chl Apr.1-Oct.15) / ∑ (number of days with valid Chl).
chl_mean_gr = mean Chl concentration during production season (missing Chl ignored):
∑ (valid daily Chl between prod_day1 and prod_daylast) / (total number of days with valid production)
chl_max = maximum Chl concentration over the year
sst_mean = same as chl_mean but then for sst
sst_mean_gr = same as chl_mean_gr but then for sst
ow_days = number of open water days between April 1 and October 15
∑ (daily open water fraction Apr.1-Oct.15)  
So a day with open water concentration of 50% will count as 0.5 days