Although there is only a limited amount of time where the processed data for all 4 benchmark met stations overlap, it is instructive to observe the charts for other combinations of overlap between stations. Thus, charts from 1990-1998 were made regardless of overlap between certain stations. It is also perhaps more revealing to observe the ‘clean’ graphs, i.e., graphs made from data sets with the flagged point removed.
The elevations of the met stations are as follows:
PRIMET: 430 m (1321 ft)
CENTRAL: 1018 m (3339 ft)
VANILLA: 1273 m (4175 ft)
UPLO: 1294 m (4244 ft)
The most curious observation is that precipitation for PRIMET exceeds that for VANILLA consistently, year after year, almost without exception even down to individual precipitation events (an obvious exception to this is the 1990 graph, where on July 5, VANILLA received 24 mm of precip compared to PRIMET’s 10.2 mm, and the cumulative graph shows VANILLA above PRIMET for the rest of the year.) This is surprising, given the fact that VANILLA is almost 3000 ft higher than PRIMET.
VANILLA’s precip is also less than CENTRAL’s, at least during the limited amount of time of overlap between those two stations’ data. Thus, the second-to-highest elevation station appears to consistently receive less precip than the two lower ones. It would be interesting to compare the precip data between VANILLA and CENTRAL over a longer period of time, but their overlap only spans about 5 months. When up-to-date processed data is acquired, this comparison will be made. CENTRAL consistently gets more precip than PRIMET, which is to be expected.
The highest elevation station, UPLO (both the stand alone and shelter gauges), consistently shows the greater amounts of precip. This is not surprising. Also, the shelter gauge catches less precip than the stand alone, which is to be expected.
All of this suggests that the seasonal precip differences between VANILLA and the lower stations (PRIMET and CENTRAL) is in fact not an undercatch problem, but a real meteorological anomaly that is yet to be explained. The differences are more pronounced during winter when large, frontal precip events inundate the region; more localized convective precip events during the summer tend not to show the anomaly. Further study to determine the causes of this phenomenon (i.e., certain types of storm/temp/precip scenarios) will be eventually be done.
The different types of precip events (storms) affecting the HJ Andrews were also studied in order to ascertain whether different event characteristics might affect the trends between upper and lower sites. Days were taken during storm periods and stratified by temperature, wind speed/direction, cold vs. wet storms, and snow vs. rain storms. Windy storms accompanied by cold temperatures would likely result in undercatch at the upper stations, leading us to question the validity of these stations’ precip sensors.
Results of this study can be found under ‘Investigation of particular precipitation events’. Results are not strongly indicative of an undercatch problem, though the charts hint that an undercatch scenario may be partly to blame for the higher precip amounts at PRIMET. I suspect that there is a real meteorological phenomenon causing higher precip at PRIMET which is sometimes exacerbated by a slight undercatch problem, but the undercatch problem is not the cause of the phenomenon.
Without updated processed data for all the stations, it is difficult to come to any conclusive decisions regarding this question. Hence, this study will be continued when the processed data is updated.