Sorted the factor levels in the order required in the legend to avoid problems later. Reorganized ggplot calls.

Moved the fix to the right place. Modified it to be a onliner. Commented it. Added 0 instead of NA to the missing columns. Included the same fix for pct2 (extremes).

If there is a unique forecast, extreme limits and tercile limits must be vectors. Now 1-D array is cot accepted. Observations are proven to be a vector or a numeric value.

Corrected bug in labels of tercile/extreme probabilities. Now uses first panel positions for all labels.

Corrected many style changes and corrected documentation according to reviewer comments. Author name was corrected.