MarcC":1megnvb9 said:
Our weather also usually occurs in 4-6 week cycles, with the greater bulk of winter storms starting in mid-November. Project that out and you get a often dry mid-January.
This is absurd. Weather can just as easily occur in one or two week cycles. If 4-6 weeks had sufficient persistence to make mid-January abnormally dry that would show up in negative month-to-month correlations and even (horror of horrors!) in averages.
MarcC":1megnvb9 said:
conveniently forgetting that if Jan 11 had three 100" snowstorm years and zero inches in the other 28 years, that day would still have an average of 9.6". This is what I mean by saying that averages smooth out the differences and why relying just on averages is a flawed analysis.
The above is unquestionable correct. However, a tendency toward 100 inch dumps and extended dry spells vs. more consistent snowfall would show up in higher standard deviations, increased probability of >90 and <30 inch snow months, etc. This is easily evident when you compare Sierra areas with 400 inch average snowfalls to more consistent areas like Targhee, Brighton/Solitude, Wolf Creek, Whitewater.
The Alta Guard data goes back to 1946. Monthly snowfall from smallest to largest instead of chronological.
Dec Jan Feb Mar
12.3 1 34 23.8
17 19.5 35 33
17 26 39 35
20 28.5 39.75 35
27 38.5 40 44
34 39 41 45
39.5 41.8 45 45
40 42 49.7 46.5
43.1 42 50 47
47.5 44 53 47.2
49 46 53.4 50.1
50 50.5 56 53
51.5 54 56.3 53
53 56 57 53.8
53 57 58 60
53 58 58 60
55 59 60.5 61
57 61 61 62
57.2 64 61.5 63
62 64.5 61.5 63.5
63 64.5 62 64.75
64.85 66.2 63 68
65 70.75 66 69
66 73 66 70
67 73 67 70
69 74.3 68 ---
70 74.5 68 70
71 75.5 69 71
71.5 78.5 72 79
73.5 81 72 80
73.8 82.8 73 82
78.7 83.5 73.5 82
79.5 84.5 73.5 84
80 85 75 84
81.8 86 77 84.5
82.5 86 77.9 85
83 88.5 79.5 85
84 94.5 82 85
86.1 96 84 87
86.4 99.5 84.1 92
91 100 85 92.3
91 100.9 85 93
96 103 85.5 93
97 103.5 86 93
101 104 88 93.8
105 104.5 91 97
106.5 105.1 91 97
107 105.3 92.5 99.5
107 107.5 97.5 100
112 108 97.5 100
114 112 98 101
119.2 112 100.5 104
122 113 102.9 109
124.5 113 103 110
128.2 113.5 104 110.9
131 115 104 113
132 122.7 105 114
132.6 128.9 107 115.5
137 132 107 123
141 133 110 124
142 134 112.5 128.9
146.5 141.5 112.7 129
151 143 119.5 135
156 143 129 142.2
159 148 130 150
160 150 131.5 151
161 165.3 132 153.6
164.8 168 134.05 163
165 186 148 164
184 187 155 165
244.5 199.7 156.6 183
91.0 91.7 82.6 88.9
45.7 41.5 29.4 36.0
Averages and standard deviations at the bottom. Looking at the extreme lowest months at the top of the table you get a message consistent with the standard deviations: December is the most volatile month, followed by January, March and February.
MarcC":1megnvb9 said:
3. Dry does not mean there must be an inversion. He tried that flawed analysis, too.
Uhh, no! The Utah residents were the ones using the inversions as the demonstration about dry spells and socal and I who produced the SLC inversion data showing peak "inversion season" lasting from mid-December to about the first week of February. Furthermore Jim Steenburgh has explained in great detail the reasons for the mid-winter concentration of inversions.
MarcC":1megnvb9 said:
I'm more than happy to go with personal experience and local weather geeks.
My personal experience is concentrated in mid-March and has also seen an unusual number of dry spells. Since inversions are essentially non-existent by then, it was obvious to me that "dry" was a necessary but not sufficient conditions for an inversion. This eventually led to my investigation of the more recent Alta daily data and sequence of consecutive dry days in the 31 day months:
Month 1: 11,10,9,9,8,8,8,6,5,5,5,4,4,4,4,4,3
Month 2: 11,9,8,8,7,7,6,6,6,5,5,4,4,4,4,4,4,3,3,3,3
Month 3: 12,11,10,9,9,8,7,6,5,5,5,4,4,4,3,3,3,3
This did not surprise me at all because of:
1) My personal March ski experience in Utah, which dates farther back than to when either MarcC or admin (or Evren I suspect) moved there, and
2) My overall collection of ski data, in which many of the "bad" seasons (76-77, 80-81, 86-87, 11-12) are characterized by widespread December western droughts that severely retard the time at which ski areas get adequately covered.
As for "local weather geeks," I would be astounded if Steenburgh, Trumper, Thayer, any of those guys would say dry spells are more likely in January than December.
Evren":1megnvb9 said:
We know about confirmation bias and "availability heuristic" (though I had to look up the name on that one). There is also a similar one where people tend to give more weight to outlier events, because they are easier to recall.
I have mentioned before that Utah residents are more likely to recall January dry spells when they are freezing and choking on particulate smog than March dry spells when they are enjoying 60+ degree shirtsleeve weather and a choice of outdoor summer activities as well as skiing.
Recent years are always remembered more than more distant ones. Chris Steiner of zrankings.com wrote up a nice analysis of the most recent terrible western ski season. But there was no comparison or even mention of 1976-77, somewhat understandable since that was before he was born.
Evren":1megnvb9 said:
You flip a coin 40 times and it comes up tails. What are the probabilities on the 41st? How would you bet?" We all said 50-50 and it doesn't matter -- thinking we had avoided a simple curveball. He said, "No, what you have here has got to be a loaded coin. Bet on tails." It is good to have some of that in your thinking.
What we have in this situation is the last 5 years when Alta's January snowfall has been 55% of average and thus by far the driest winter month. Even the last 15 years January has been 86% of long term average and lower than both December and March. You push that data back
40 years and the presumed effect disappears. Volatility of snowfall is such that a decade is marginally adequate for estimating season snowfall. For individual month snowfall patterns I would think 20 years would be the bare minimum to consider drawing this kind of conclusion.