...
dailyTime, dailyAverage, dailyCount are copies of time, average, count in the normal 'data' structure, with data gaps filled. The code is capable of calculating these variables from hourly or other averaged data (currently we use daily data, so this step is essentially a pass-through). The dailyAverageByYear, dailyCountByYear are a reshape of the dailyAverage and dailyCount data into a matrix that is year by date, where the year is specified by iYear and the date is 1 to 365 (leap days excluded), so the matrix is n years by 365 days. dayOfYearAverage, dayOfYearCount, dayOfYearStdDev, dayOfYear33Percentile, dayOfYear66Percentile are the day of year statistics for each of the 365 years of the year. The , calculated using the dailyAverageByYear data only, without weighting by dailyCount or recalculation from the raw data. This average of an average approach is normal metrological procedure, as many metrological data sets only report daily min/max average values. The dayOfYearCount is the count of the raw readings on that day of the year, whereas dayOfYearCountDays is the number of days that have data. So for the 15 years of Saanich Inlet, if only 5 years have data on a specific day, that dayOfYearCountDays value is 5, but dayOfYearCount will be ~400000 readings. dayOfYearCountDays is used to determine if there is enough data (minimum 3) for a valid percentile calculation. dailyAverageAbove66Percentile, dailyAverageBelow33Percentile are the values used the anomaly plot, so that they are daily time series of the same size as the daily average data, and their values are either the dailyAverage or NaN if they are not above/below the percentile. The anomaly plot is then simply:
...