USGS models show a 50% decrease in Conus snowfall by the 2060s

[The_Global_Warmer]

http://www.drroyspencer.com/2014/05/u-s-snowfall-to-decrease-by-50-in-50-years/

Most effected regions are the inter mtn west, the Great Lakes, and the Northeast.

[ORH_wxman]

Its hard to take models seriously that can't even hindcast the past...this is laughable.

 

 

 

 

 

 

 

 

Not only that. The models thinks we should have lost over 10% of annual snowfall since just the 1980s.

 

 

 

Actual (this is over ORH where it predicts larger drops than the most of the state):

 

 

 

 

 

 

 

 

Climate models forecasting a paramater like snow...especially on a regional level....is essentially worse than just predicting climo. But they are even terrible on larger scales...such as they have been predicting a big decrease in winter snow cover over the N Hemisphere and we've seen the exact opposite. The biggest reason is probably because the models warm us way too quickly versus reality.

[GaWx]

It seems like the AGW alarmists are covering their bases. I seem to recall that some of them were actually calling for an INCREASE in snowfall lol. Now, if any of them are calling for little change, they'll have all bases covered and could say "I told you so" regardless of the outcome.

Edit: For those confused or who don't actually read the article, Dr. Spencer, himself, is not at all in the AGW alarmist camp. He's noting a model projection that he thinks is hogwash/laughable.

[The_Global_Warmer]

It seems like the AGW alarmists are covering their bases. I seem to recall that some of them were actually calling for an INCREASE in snowfall lol. Now, if any of them are calling for little change, they'll have all bases covered and could say "I told you so" regardless of the outcome.

Edit: For those confused or who don't actually read the article, Dr. Spencer, himself, is not at all in the AGW alarmist camp. He's noting a model projection that he thinks is hogwash/laughable.

 

 

1.  It's just a application tool from the USGS.

 

2.  It's based on RPC 8.5 which is a very high Co2 projection IIRC.

 

 

I expect snow loss will happen as we continue to warm.  I doubt it will be that bad by the mid 21st century.

[nflwxman]

Regional projections are tough especially in time periods as short as 2 decades.  As we all know, snowfall climatology is quite chaotic, even over a 20 year period, which inherently makes it even harder to model than temperatures or RH, IMO.

 

The two ways we could potentially improve this type of modeling is:

 

1) Higher resolution inputs.  I think regional modeling falls victim to grid scale inputs that do not factor great elevation parameters (for example).

 

2) Evaluate the model on longer time periods- Since snowfall climo is so chaotic, these models probably can't be evaluated on anything less than a 30 period time period.  

 

ORH, I would not necessarily discount the long term viability of this model based on a 20 year snippet in the state of MA.  I think many here expect lock step decade by decade real world vs. climate models, even though they are not really designed for short term evaluation.  Also, I'd be curious if the entire NE hindcast performs better than just MA.

 

I agree though, that large scale snowfall models so far have not been spot on yet (for various reasons).  Just playing devil's advocate here.

[ORH_wxman]

Regional projections are tough especially in time periods as short as 2 decades.  As we all know, snowfall climatology is quite chaotic, even over a 20 year period, which inherently makes it even harder to model than temperatures or RH, IMO.

 

The two ways we could potentially improve this type of modeling is:

 

1) Higher resolution inputs.  I think regional modeling falls victim to grid scale inputs that do not factor great elevation parameters (for example).

 

2) Evaluate the model on longer time periods- Since snowfall climo is so chaotic, these models probably can't be evaluated on anything less than a 30 period time period.  

 

ORH, I would not necessarily discount the long term viability of this model based on a 20 year snippet in the state of MA.  I think many here expect lock step decade by decade real world vs. climate models, even though they are not really designed for short term evaluation.  Also, I'd be curious if the entire NE hindcast performs better than just MA.

 

I agree though, that large scale snowfall models so far have not been spot on yet (for various reasons).

 

 

I don't expect it to be perfect...but this projection is downright laughable.

 

The reason the projection is terrible is because the warming it shows is not realistic. The warming trend of the Northeast U.S. in winter time is 0.3F per decade since the early 20th century. In order to reach the temperature projections of those models, we would need to warm at roughly 1.5F per decade between now and 2060. It shows a warming of around 6-7F in the next 40-50 years. (about 8F warmer than the 1950-2005 mean)

 

If you cherry pick the most extreme 30 year period of temp rises in the past observations, you could come up with that type of warming trend. That occurred between both 1977-2007 and 1903-1933 which warmed at 1.6F per decade in the Northeast. However, that trend of course did not keep up and our most recent 30 year trend from 1984-2014 is +0.5F per decade and the trend is actually negative since the mid 1990s.

 

I don't find it credible that we will "all of the sudden" jump to an average rate 3 times higher than the previous 30 years over the next 40-50 years.This isn't just a Northeast phenomenon either, its pretty much in any region you look at in the US. Its actually worse in most other regions because the Northeast has seen less cooling recently than many other parts of the U.S. in the past 20 years.

[Jonger]

1.  It's just a application tool from the USGS.

 

2.  It's based on RPC 8.5 which is a very high Co2 projection IIRC.

 

 

I expect snow loss will happen as we continue to warm.  I doubt it will be that bad by the mid 21st century.

 

Maybe, maybe not... 

 

You are either on the right side of the low or you aren't at this point. Winter temps in the US have decreased compared to the 1990's, so we aren't heading the right direction for that projection to verify at this point.

[skierinvermont]

It's just because it's using RCP8.5. 

 

RCP 4.5 probably shows a more realisitc loss. Although it might not be a regional model best suited for this purpose.

 

A 20-40% loss in snowfall in SNE by the 2060s seems reasonable to me. 

[ORH_wxman]

It's just because it's using RCP8.5. 

 

RCP 4.5 probably shows a more realisitc loss. Although it might not be a regional model best suited for this purpose.

 

A 20-40% loss in snowfall in SNE by the 2060s seems reasonable to me. 

 

 

RCP 4.5 is essentially lock and step with 8.5 until around 2050. That's when most of the loss in snowfall occurs.

 

I don't find it credible. Especially given how far too low it is already in the 2000s...the true starting point would probably mean a reduction of at least 50-60% by 2060 rather than their modeled 2000s starting point.

 

But even if we ignore the poorly modeled starting point for snow, the RCP4.5 shows 3.5F of wintertime warming in SNE by 2037. (the 2025-2049 mean temp) That still doesn't come across as very realistic.

[michsnowfreak]

This is laughable on so many levels. Like, more than GFS 384hr laughable. Back in the 1990s there was an article that said something like local (MI) snowfall would decrease like 50% by 2020 or 2030 (they compared 2020s Michigan winters to 1990s Kentucky or southern Ohio winters, so actually that may be more than 50%). How is that trend working out?

 

Snowfall is increasing...and dramatically....in the Great Lakes so far in this 21st century. It is actually to the point where a reverting back to where we were in the mid-20th century would be a huge slap in the face to us winter-lovers, LET ALONE decreasing 50% in 50 years from 1950 levels.

 

Looking at Detroit snowfall alone (I do not have other cities decadal averages handy, but Im sure the trends are similar) tells the tale...and unlike the east coast, we do not live or die by the big storm. It shows that snowfall it every facet of the game (total, frequency, big storms) is increasing.

 

Detroit, MI snowfall avg by decade:

1900s- 46.3"

1910s- 39.7"

1920s- 46.1"

1930s- 32.9"

1940s- 27.6"

1950s- 37.8"

1960s- 31.8"

1970s- 45.6"

1980s- 45.2"

1990s- 37.2"

2000s- 45.3"

2010s- 59.4"

 

Will..how did you do that graph for MA? Id like to do one for MI.

[ORH_wxman]

This is laughable on so many levels. Like, more than GFS 384hr laughable. Back in the 1990s there was an article that said something like local (MI) snowfall would decrease like 50% by 2020 or 2030 (they compared 2020s Michigan winters to 1990s Kentucky or southern Ohio winters, so actually that may be more than 50%). How is that trend working out?

 

Snowfall is increasing...and dramatically....in the Great Lakes so far in this 21st century. It is actually to the point where a reverting back to where we were in the mid-20th century would be a huge slap in the face to us winter-lovers, LET ALONE decreasing 50% in 50 years from 1950 levels.

 

Looking at Detroit snowfall alone (I do not have other cities decadal averages handy, but Im sure the trends are similar) tells the tale...and unlike the east coast, we do not live or die by the big storm. It shows that snowfall it every facet of the game (total, frequency, big storms) is increasing.

 

Detroit, MI snowfall avg by decade:

1900s- 46.3"

1910s- 39.7"

1920s- 46.1"

1930s- 32.9"

1940s- 27.6"

1950s- 37.8"

1960s- 31.8"

1970s- 45.6"

1980s- 45.2"

1990s- 37.2"

2000s- 45.3"

2010s- 59.4"

 

Will..how did you do that graph for MA? Id like to do one for MI.

 

 

Microsoft excel...if you have seasonal snow totals for Detroit, it would probably take about 2 minutes to make that graph. Just plot them on a bar graph (or line if that is your preference) and run a smoothing filter (I usually do like 8-10 year running average to smooth it out).

 

 

I obviously agree with you on the laughable part of those projections. Its hard to take a climate model seriously that turns winter in Boston into a DC climate within 4-5 decades. That's really in a nutshell what is wrong with climate science right now...we don't have very accurate models for a lot of sensible wx that people want to know about...that includes temperature too.

[nflwxman]

Microsoft excel...if you have seasonal snow totals for Detroit, it would probably take about 2 minutes to make that graph. Just plot them on a bar graph (or line if that is your preference) and run a smoothing filter (I usually do like 8-10 year running average to smooth it out).

 

 

I obviously agree with you on the laughable part of those projections. Its hard to take a climate model seriously that turns winter in Boston into a DC climate within 4-5 decades. That's really in a nutshell what is wrong with climate science right now...we don't have very accurate models for a lot of sensible wx that people want to know about...that includes temperature too.

 

I wonder if these models consider the Francis hypothesis of the lagging jet stream.  I know a few skeptics here have reservations about that theory, but circumstantial evidence suggests that could be big part of the uptick in eastern US snowfall in the last decade or two. It's been very feast or famine.

[ORH_wxman]

I wonder if these models consider the Francis hypothesis of the lagging jet stream.  I know a few skeptics here have reservations about that theory, but circumstantial evidence suggests that could be big part of the uptick in eastern US snowfall in the last decade or two. It's been very feast or famine.

 

 

Its possible...but there's also a bit of a cycle too due to the natural variability in the NAO on decadal timescales. Determining how much to weight each one is obviously an area for further research.

 

The snowfall isn't the only issue though. The temperatures are a problem that don't really pass the "smell test" either. A lot of the snowfall decrease is because the models think temperatures are going to warm about 4 times faster than they have in the past 30 years.

[WXheights]

I wonder if these models consider the Francis hypothesis of the lagging jet stream.  I know a few skeptics here have reservations about that theory, but circumstantial evidence suggests that could be big part of the uptick in eastern US snowfall in the last decade or two. It's been very feast or famine.

totally agree but just not there yet as complete proof.

 

Its possible...but there's also a bit of a cycle too due to the natural variability in the NAO on decadal timescales. Determining how much to weight each one is obviously an area for further research.

 

The snowfall isn't the only issue though. The temperatures are a problem that don't really pass the "smell test" either. A lot of the snowfall decrease is because the models think temperatures are going to warm about 4 times faster than they have in the past 30 years.

Variability with possible giant snowfalls and other years very little. Remember nothing linear.

 

Eventually lower elevations -- southern New England all done except under special rarer regimes - further north might be snow gains at summits downward trends base lodge and lower.

 

Small data set but worthy. http://www.uvm.edu/~waw/archives/Mansfield%202009.pdf

[ORH_wxman]

totally agree but just not there yet as complete proof.

 

Variability with possible giant snowfalls and other years very little. Remember nothing linear.

 

Eventually lower elevations -- southern New England all done except under special rarer regimes - further north might be snow gains at summits downward trends base lodge and lower.

 

Small data set but worthy. http://www.uvm.edu/~waw/archives/Mansfield%202009.pdf

 

 

Under what scenario and what timescale though?

 

I am very skeptical of a 7F increase by 2060 here and a 3.5F increase by 2040. Remember, Southern New England has seen no trend in snowfall since the middle 20th century. There's actually been a slight (albeit insignificant) positive trend since the early 20th century (we had some dud years in the 20s/30s). This is all of the sudden going to turn into a 50% decrease in the next 4 decades? Never say never, but I'll certainly take the under on that percentage number.

 

At least this hypothesis is testable...though we'll have to wait a few decades.

[WXheights]

Under what scenario and what timescale though?

 

I am very skeptical of a 7F increase by 2060 here and a 3.5F increase by 2040. Remember, Southern New England has seen no trend in snowfall since the middle 20th century. There's actually been a slight (albeit insignificant) positive trend since the early 20th century (we had some dud years in the 20s/30s). This is all of the sudden going to turn into a 50% decrease in the next 4 decades? Never say never, but I'll certainly take the under on that percentage number.

 

At least this hypothesis is testable...though we'll have to wait a few decades.

Sorry I was thinking end of century. Mid century 2060 maybe good for about 30-40% reduction of snowfall - even more in depths though.

[Jonger]

Let's say climate sensitivity is in the upper range.... You really think we are going to let things get that out of hand before going to geo-engineering?

I'll be shocked if we hit our first GISS year 1.25C before 2030. It'd hard to say, but at the current pace it looks doubtful.

[OSUmetstud]

Snowfall is just about the least accurate thing we measure in meteorology, even snowstorm to snowstorm.  How much of the seeming increase in Great Lakes snowfall is due to the implementation of the universal 6 hr snowfall measurement technique?  Is New England snowfall on the "decrease" but is being covered up by the newer guidelines?  

 

Areas of higher snowfall to water ratio snows, like the Great Lakes, northern New England, and the Intermountain West would seem the most susceptible to the guidelines.  

[JC-CT]

Snowfall is just about the least accurate thing we measure in meteorology, even snowstorm to snowstorm. How much of the seeming increase in Great Lakes snowfall is due to the implementation of the universal 6 hr snowfall measurement technique? Is New England snowfall on the "decrease" but is being covered up by the newer guidelines?

Areas of higher snowfall to water ratio snows, like the Great Lakes, northern New England, and the Intermountain West would seem the most susceptible to the guidelines.

Excellent point. As long as a certain poster remains in Tolland, CT, reported snowfall will continue to rise indefinitely.

It stands to reason that snowcover would decrease much more rapidly than snowfall, given the difference between 5 snowstorms vs. 4 snowstorms plus 1 pack-destroying rainer. It doesn't affect SNE that much compared to the ski industry of NNE.

[The_Global_Warmer]

Global warming has already radically altered snow cover in my area.  Not snowfall yet.  But snow cover has been changed a lot.

 

 

This past winter was the snowiest in terms of snowfall ever in my lifetime going back to 1982 in my city.  40" fell.  Now while this year snow cover was better than any year since 2000 here.  Even 09-10 when we had 30",

 

If only the data wasn't almost non existent.  But this is something that has been observed between the 35-40N latitude belt East of the Rockies.  It gets worse the further South you go obviously.

 

 

I totally get why snow lovers further North are so worried about their snow being hosed by AGW.  I remember vividly realizing this wasn't going to stop and only get worse. 

 

Like temps this varies.  I would assume this phenomenon will keep creeping North.

 

while only getting worse further South.

[Jonger]

Snowfall is just about the least accurate thing we measure in meteorology, even snowstorm to snowstorm. How much of the seeming increase in Great Lakes snowfall is due to the implementation of the universal 6 hr snowfall measurement technique? Is New England snowfall on the "decrease" but is being covered up by the newer guidelines?

Areas of higher snowfall to water ratio snows, like the Great Lakes, northern New England, and the Intermountain West would seem the most susceptible to the guidelines.

I live in Michigan, the 1990's were pretty bad for synoptic snow storms. So the increase in the 2000's and 2010's isn't an illusion, I have been monitoring snowfall for 20-25 years in SE Michigan. High ratio snow is lake derived, this would barely affect the Detroit statistics.

[JC-CT]

Global warming has already radically altered snow cover in my area. Not snowfall yet. But snow cover has been changed a lot.

This past winter was the snowiest in terms of snowfall ever in my lifetime going back to 1982 in my city. 40" fell. Now while this year snow cover was better than any year since 2000 here. Even 09-10 when we had 30",

If only the data wasn't almost non existent. But this is something that has been observed between the 35-40N latitude belt East of the Rockies. It gets worse the further South you go obviously.

I totally get why snow lovers further North are so worried about their snow being hosed by AGW. I remember vividly realizing this wasn't going to stop and only get worse.

Like temps this varies. I would assume this phenomenon will keep creeping North.

while only getting worse further South.

We will just have to move north where the economy will be booming due to the new exploitation of previously unaccessible oil and gas fields.

[Jonger]

Global warming has already radically altered snow cover in my area. Not snowfall yet. But snow cover has been changed a lot.

This past winter was the snowiest in terms of snowfall ever in my lifetime going back to 1982 in my city. 40" fell. Now while this year snow cover was better than any year since 2000 here. Even 09-10 when we had 30",

If only the data wasn't almost non existent. But this is something that has been observed between the 35-40N latitude belt East of the Rockies. It gets worse the further South you go obviously.

I totally get why snow lovers further North are so worried about their snow being hosed by AGW. I remember vividly realizing this wasn't going to stop and only get worse.

Like temps this varies. I would assume this phenomenon will keep creeping North.

while only getting worse further South.

Do you have a chart showing snowfall by the decade in STL? It might be more of a case of bad luck. Detroit just had it's snowiest 15 year period of all time, while Toronto had a pretty weak period. Toronto has the same average weather as Detroit. It was just bad luck in their case.

Top 25 years of snowcover years in Detroit is peppered with 2000-2014 years.

[ORH_wxman]

Snowfall is just about the least accurate thing we measure in meteorology, even snowstorm to snowstorm.  How much of the seeming increase in Great Lakes snowfall is due to the implementation of the universal 6 hr snowfall measurement technique?  Is New England snowfall on the "decrease" but is being covered up by the newer guidelines?  

 

Areas of higher snowfall to water ratio snows, like the Great Lakes, northern New England, and the Intermountain West would seem the most susceptible to the guidelines.  

 

 

The 6 hour measurements really only affect larger storms (say 15"+)...the difference is pretty negligible under those amounts and thus the seasonal amounts. We know this by looking at coops that only measure once per day and their snowfall trends are pretty much in line with first order stations or known 6 hour coops. Though granted, good coops are hard to find as data for snowfall tends to be choppy on them...with inconsistency in the number of missing dates.

 

I agree that if there was an area where the 6 hour guidelines affected the trend the most, it would be those very high ratio areas. But for trendlines in areas like BOS and ORH, they don't matter.

[JC-CT]

The 6 hour measurements really only affect larger storms (say 15"+)...the difference is pretty negligible under those amounts and thus the seasonal amounts. We know this by looking at coops that only measure once per day and their snowfall trends are pretty much in line with first order stations or known 6 hour coops. Though granted, good coops are hard to find as data for snowfall tends to be choppy on them...with inconsistency in the number of missing dates.

I agree that if there was an area where the 6 hour guidelines affected the trend the most, it would be those very high ratio areas. But for trendlines in areas like BOS and ORH, they don't matter.

Six hour measurements definitely matter when it comes to general public reports. That's 3 more opportunities per day to slant-stick.

[ORH_wxman]

Global warming has already radically altered snow cover in my area.  Not snowfall yet.  But snow cover has been changed a lot.

 

 

This past winter was the snowiest in terms of snowfall ever in my lifetime going back to 1982 in my city.  40" fell.  Now while this year snow cover was better than any year since 2000 here.  Even 09-10 when we had 30",

 

If only the data wasn't almost non existent.  But this is something that has been observed between the 35-40N latitude belt East of the Rockies.  It gets worse the further South you go obviously.

 

 

I totally get why snow lovers further North are so worried about their snow being hosed by AGW.  I remember vividly realizing this wasn't going to stop and only get worse. 

 

Like temps this varies.  I would assume this phenomenon will keep creeping North.

 

while only getting worse further South.

 

 

North America has seen an increase in snow cover though.

[OSUmetstud]

I live in Michigan, the 1990's were pretty bad for synoptic snow storms. So the increase in the 2000's and 2010's isn't an illusion, I have been monitoring snowfall for 20-25 years in SE Michigan. High ratio snow is lake derived, this would barely affect the Detroit statistics.

This is somewhat true, but not all the way true.  Even Detroit's synoptic snowfall is on average drier compared to let's say Boston or NYC.  

[OSUmetstud]

My point was mostly to spur a discussion, I didn't think all of the increase in Great Lakes snowfall was due to the 6 hr measurement technique, but I think some of it could be.  

[michsnowfreak]

totally agree but just not there yet as complete proof.

 

Variability with possible giant snowfalls and other years very little. Remember nothing linear.

 

Eventually lower elevations -- southern New England all done except under special rarer regimes - further north might be snow gains at summits downward trends base lodge and lower.

 

Small data set but worthy. http://www.uvm.edu/~waw/archives/Mansfield%202009.pdf

While thats feasible for New England...as mentioned earlier...the Great Lakes region is one of more frequent snowfall rather than big storms/feast/famine. So if anything you would expect a lessening of the small snowfalls here....and thats not happening. Both big and small snows are increasing.

[michsnowfreak]

Snowfall is just about the least accurate thing we measure in meteorology, even snowstorm to snowstorm.  How much of the seeming increase in Great Lakes snowfall is due to the implementation of the universal 6 hr snowfall measurement technique?  Is New England snowfall on the "decrease" but is being covered up by the newer guidelines?  

 

Areas of higher snowfall to water ratio snows, like the Great Lakes, northern New England, and the Intermountain West would seem the most susceptible to the guidelines.  

What do you mean the implementation of the 6-hour rule? I asked a met (Bill Deedler) from DTX if he knows how long they have been doing 6-hour increments and he said as for as far back as he knows, probably the beginning of record. (He worked at the NWS from 1974-2011, so at least that long). If any period of snowfall record is tainted with snowfall I would say its the FAA years from the mid-1990s to mid-2000s when NWS offices had to beg the FAA to do half-a** snow measurments until NWS employees or snow paid observers took over. Those years many stations basically have low-balled data. I have seen in old data many times when snowdepth the next day would be lower the next day than the snowfall, indicating settling.