psuhoffman

Moving this here so as not to clutter up the main thread any further. Counter points... 1) Using grades might not be the best example since my district enforces a minimum grade of 50 so as not to skew students grades with 0s lol. But lets not get into that please. Thought I would start with some levity. 2) The claim that outliers don't skew the mean is predicated on the mean of the whole dataset. But we are comparing the change in mean over time. To do that you have to chunk the mean into smaller portions. Climo uses a 30 year mean. If we remove just 4 years out of those 30 the current mean for DCA is only 9". Those 4 years are most definitely skewing the mean of the current 30 year period...and by doing that is also skewing the slope of the trendline. 3) This depends what you are trying to tease out of the data. The mean tells you over a period of time how much total snowfall is likely to fall proportional to each year. But it does not tell you how much snow is most likely to fall in any given year. The truth is over the last 30 years that 9" number (what you get if you remove the 4 super outliers) is way closer to what you should expect going into any given year than the 14" mean number with which was only reached 26% of the time. In any given season there was a 74% chance you were getting less snow than that. If you are trying to calculate how much snow you should expect in any given season, the mean is not the best tool. 4) Maybe we are trying to get at different things. What I am getting at are in any given season the odds of getting a specific amount of snow are dropping. The odds of getting 8", 10", 12", 15" and so on...going into any season are going down. And going down a LOT. The odds of getting a 4", 6", 8" storm are all dropping. That is what I am focused on. That going into any specific season the chances of it being crap are going up. The mean does not capture that fact as well. 5) You keep focusing on what the change is per year. Over time very small changes add up to be significant. 6) In a vacuum its correct to say 99% is attributed to random chance because the math you are using doesn't know any corroborating evidence. But we do know some additional facts. The fact that it is getting warmer for instance. If you apply logic to the fact that at the same time the slope of snowfall is decreasing correlates to an increase in temperature the amount we should attribute to "random chance" goes down. 7) Lastly I am not making this about human caused climate change. Maybe this is cyclical. Maybe its not human caused. Maybe at some unknown point in the future the trend reverses. Those are all true. But that is not what I am getting at. I am not living at some random unknown point in the future. I don't care what the chances of snow will be is 2323 I care what the chances of snow are right now in 2022. For the sake of this argument I don't care if the warming is man made or natural it is still impacting my snowfall right now just the same. And since we don't know when or if the trend will reverse I don't even care about that. What my analysis was showing is that our chances of snow in any given season right now are significantly lower than they were during recent recorded history. It's not even a predictive thing. I am not making a statement about what snow probabilities will be 50 years from now. I have no idea. I am making a statement on what they are RIGHT NOW. Your analysis isn't wrong. Everything you say is 100% true. But you are focusing on different metrics that me and teasing different things out of the data. You seem to be more focused on predictive measures of snowfall over any long period of time with statistical certainty. That is a very different concept than trying to calculate what the probability of a given amount of snow in one specific season is right now and how that probability has changed over the last 100 years. We are looking at very very different things.

I left El Niño’s out when I posted the stats because they more commonly start warm then flip snowy. I posted all neutral and Nina years. La Nina’s are more often front loaded so when they start bad it can get really bad. However, we’ve had a bit of a Nino ish pattern so…if this year doesn’t behave like a typical Nina it leaves the door open more. Also as Ji pointed out most of the past Ninas that were snowless were warm early. Nina’s that start very warm tend to stay warm. But it was kinda cold just no snow. There are good reasons to cling to hope. But I wouldn’t use El Niño years since they are notorious for flipping in January.

We have a below average temp month this December ...we had a good pattern. The no snow was a fluke.....its different from other shutout December months I tend to think there is some validity to this. I commented a few days ago that perhaps the warning has made a snowless start more common also making a turn around more likely. Early season seems to be most impacted. But I’m just posting statistics. I’ve always done this. Good or bad. Ironically many years ago I was on the flip side of this and I remember some including Mitch being frustrated at me for being too optimistic! Truth is I bust high way more than low on snowfall. I’m not pessimistic. It’s just been so god awful recently it makes it seem that way when I simply post the objective statistics.

Fringed

We should be discussing snow threats…oh wait

maybe read some before posting

To comfort those on the ledge...its still extremely likely it snows at some point this winter, even if not very much. We aren't at the point yet where its likely we get a complete shut out.

I love snow...and I REALLY hope is snows 100" the next nino winter. But the evil me kinda takes pleasure knowing at least if it is a total fail I get to see the total nuclear meltdown on here.

We just had a raging -AO, about as negative an AO as we can ever get...we just didn't get any snow out of it lol.

I didn't claim my stats proved anything...I simply laid out facts and explained why I disagreed with his analysis of the data. Anyone is welcome to disagree with me. I laid out the data, people can decide for themselves which interpretation they agree with.

I will enjoy any snow we get. If we don't get any I'll take my kids to Blackwater Falls in WV after they get an upslope event to play and go sledding. I will see snow a few times when I go skiing in New England or Colorado. It will be ok. But we were all hopeful maybe this would be a "good for a nina" season...where we get close to median or maybe even closer to avg if we got lucky. But it is becoming increasingly likely this is going to be one of the bad ones...like bad bad not just just kinda bad. But likely doesn't mean definitely. Even if every past comp was awful there aren't enough years to say its a statistical certainty. But we all know this is how the really awful years start. I still do have some hope...these statistics are just data. Not a prediction. But I will lay this out there...if the PV starts to couple with the TPV and we see the AO go extremely positive in January its game over.

this implies we have a peak climo anymore lol

If I make it to January 10 without 1" of snow the numbers get even uglier. It's a VERY small list...but most are years that ended with almost no snow at all around DC.

I will take you at your word that you are not trolling and engage one last time with what I take issue with in your statistical analysis. But first I'll post some data that I will reference in my response below. @Terpeast I have used only DCA data from 1942 onwards. DCA annual snowfall with a linear trendline and 30 year running mean imposed. 18 year running probability of getting at least 10" of snow in any given season 18 yeas was not arbitrary but was chosen because in this discussion yesterday 18 years was agreed upon as a minimum data set to avoid overly skewed results to get a statistically significant result. Some additional facts. The running 30 year mean snowfall at DCA has decreased from 17.8" to 13.9" over the period of record for the airport. This trend is continuing and is getting even worse...the current 15 year mean is only 13.6" The median has decreased from 15.8" to 10.85". This trend is also getting worse. The current 15 year median is 7.8" The chances of getting at least 10" of snow in a given season have gone from 72% to 50% in a given season. This trend also is getting worse...the current 15 year probability is 47% The chances of getting 15" in a season have gone from 53% to 27%. The chances of DC getting 20" have gone from 40% to 17% 1) The words bolded above are all your opinion. The slope is NOT 0, virtually is opinion...yes over any one year the change is insignificant but over a long period of time .03 adds up to 4" of snowfall we have lost annually over the period of record. That is not insignificant when your average is only about 14". That's more than 20% of our annual snowfall we have lost. 2) Single digit snowfall years is not arbitrary, its a way of showing what chance there is of getting a significant amount of snow in any given year in DC. We can use another number if you want...8", 12", 15", 18", 20"...they all show the exact same trend so it doesn't matter. I didn't cherry pick 10 to skew the data, I picked 10 because its a nice simple even number to highlight the issue which is in any given year the odds of getting snow is going down. That is true whether you use the threshold of any amount between about 7 and 20", outside that you get some crazy useless percentages because you're using a number outside a standard deviation. 3)I've repeatedly said that using the probability of snowfall or the median is way more useful to a climate like DC and you repeatedly dismiss that which is your opinion and fine but I will explain what is wrong with a mean. DC snow climo is inflated by anomalously snowy seasons like 1996, 2003, 2010 and 2014. But the snow that happens in those winters doesn't make the years in between any less awful. They do not affect what a typical winter is. Getting a huge amount of snow every 7 years or so affects the mean a lot but isn't indicative of what any given winter is likely to be like...which is probabilistically much more likely to be one of those other 6 years. Median filters out those big years better to give you a better indication of what any given specific season is likely to be like. 4) You've said there is no evidence that the mean is being skewed but the evidence is right on the chart for everyone to see. While the median and probabilities of snowfall are clearly decreasing the range of snowfall in any given year over a period of time is increasing. The standard dev of snowfall is increasing. Even though the baseline for snowfall is lower we have had several of the biggest snowfall seasons recently. This is preventing the mean from dropping as quickly as the probabilities of snow in any given season are. But for our purposes the probabilities are more important. Who cares that it snows a crap ton once every blue moon when the truth is we spend 80% of the time in the total suckage periods in between that are getting worse! 5) You repeatedly try to manipulate the data by using an arbitrary date in the 1980's from which you can say "snowfall has increased since". That is really bad statistically because snowfall does run in cycles. We are no doubt in a down cycle right now. I have never implied DC won't have better snowfall periods ahead. But by cherry picking a date that is a minimum with which to compare the current point in time fails to acknowledge that over a longer period of time its evident that the "snowy" cycles are becoming less snowy and the "dreg" periods are becoming MORE dreg. The current down cycle is no doubt a down cycle...but its worse than previous down cycles. The recent snowy periods weren't as snowy as past snowy periods. Over longer periods the downward trend is evident.

He is trolling... I should have noticed sooner. But no need to continue engaging with childish behavior.

Last time it got this salty in here I got my biggest snow of the last 6 years so I’m liking the vibe.

This X1000000. How anyone can argue with that is beyond me. Yea we can still get an anomalously snowy winter like 2010 and 2014 that skews an “average” but the odds of getting less snow in any given year are going up and I think many have been clear that’s what we’re talking about.

You’re all over the place. 1) Earlier you said “10 years isn’t a big enough sample” and now you’re basing this analysis on 10 years. 2) 1980s avg DC snow 17.1” Median 17.8 with 3 single digit seasons Most recent 10 years avg DV snow 12.3” median 10.5 with 5 single digit seasons. By what metric were you saying the 80s were worse??? 3) If what you mean is you’re comparing 1980s to 2010s then you are still dealing with the 10 year sample size issue, but then my point why I don’t care about the mean cones up. Yes the 1980s (which you cherry picked because they were a previous snow minimum) mean was slightly lower at 17.1 to 17.2. But the 2010s were severely skewed by 2010 and 2014. But the median and single digit snowfalls tell the story. The median snow in the 2010s was only 13.5” compared to 17.8” in the 1980s. There were 4 single digit seasons compared to 3 in the 80s. So even cherry picking the previously worse minimum period and comparing to the 2010s which included 2 extremely anomalous snowy winters…the odds of less snow in any given season were still higher in the 2010s compared to the 1980s which is what I’m talking about. I don’t care about the rare once in 9 years huge seasons that skew and inflate the mean. I’m talking about the fact that in any given season the odds of DC getting less snow (or a single digit season to make the concept simple) is going up and that’s a fact no matter how you try to slice up the data.

Seriously? All you have to do is look at the DC snowfall data. From 1888 to 2000 a single digit snowfall season was fairly rare, then suddenly they are happening more often than not. You want me to waste time throwing that into chart just to prove what you can see from a 10 second look at the data?

The problem has nothing to with the number of years/sample size. The issue is you aren't looking at what I am talking about. I am talking about how DC is getting more single digit snowfall seasons than before and you are looking at a mean which can be skewed by a minority of seasons to hide that phenomenon. We are two ships passing in the night. Neither of us is refuting the other...we are simply focusing on two completely different phenomenons.

Is it partly cloudy or partly sunny?

This is what I suspect. I meant the mean is low and bounded by 0 like Dallen said not that the sample size is low. I think median gives a much better idea of what a "typical" winter was like during any given period of time. For example the last 16 years (this is the period of time where I've really noticed the phenomenon I am hinting at here) there have been 9 single digit snowfall seasons, 4 years between 10-20", 1 season between 20-30" and 2 BIG years above 30" that are skewing the mean. During that time the mean is over 13" but the median is around 8". 8" being much closer to what a typical winter has been in DC over that period since 9 years were single digit snowfall and only 7 were more than that. Thank you, you said what I meant but much better. You're totally right, we need a larger sample size than 10 so I wouldn't use a 10 year median. But you could do a running 20 or 30 year median. But something is going on that is not refutable which is the prevalence of single digit snowfall seasons in DC is skyrocketing lately. Looking at the data there is no 16 year period that even comes close to having 9 single digit seasons. 16 years is starting to become a bit much to continue the "its just recency bias" or "its just a short term blip" mantras. And it's not like its a little off...its way off. I know 16 years is kinda arbitrary but going back over all previous 16 year periods DC had 5, 4, 3, 5, 5, 3, and 3 single digit snowfall seasons....yet during the most recent 16 years there were 9. Even if you try to simply cherry pick the worst 16 year periods you can't get close. The worst you can find cherry picking is another 16 year period with 6 unless you overlap the current period and then you can find 7 single digit seasons, but that is only by including several of the seasons in this current 16 year period. Basically the current most recent 16 years have been the absolute worst in recorded history in DC for getting abysmal low snowfall seasons...and the next worst is a 16 year period overlapping this current 16 year period! And nothing else comes close. That seems significant to me, more significant than to dismiss with "just random chance". The mean doesn't show it as much because a few seasons have been skewing the number enough to hide what I am talking about...which is getting more frequent dreg single digit snowfall seasons . Now what we don't know is this....is this awful stretch just part of a random cyclical thing or a more permanent shift. It's very possible that if records went back further we could find similar awful periods...and perhaps we pull out of this and in the next 16 years only have 3 or 4 single digit seasons (but do you really wanna bet on that???). So I am NOT saying I know this is all due to warming (although I suspect some of it is) but what I am saying is we are in a pretty long term stretch where its more likely than in any other period in recorded history for DC to have single digit snowfall year to year. It's been a long enough period now that until we break out of it...I don't consider it just a blip. Its not been just a few years...this trend has been going on for quite a while now with just a few interruptions by some big seasons here and there. But other than those big seasons....suddenly DC's typical winter is a single digit snowfall...when that used to relatively rare in the past.

A mean of such a low number can be skewed by a few years. I’d be more interested in a statistical analysis of the median and changes in the standard deviation.