DCA Snowfall Total Controversy

[famartin]

Huge. We are creating an important body of record that goes back 130 years.  The difference between #8 all time and #4 all time is 1.4".  

One other thing:  Really, DCA records should probably not be compared to downtown.  Just like with virtually every other major city, offices move to airports and create a discontinuous record.  Best not to squint too hard at them.  PHL set a bunch of record lows in the 40s-60s which still stand.  Why?  Because it moved to the airport from downtown.

[JonathanW]

It seems to me what's being described below is essentially random noise in the measurements. Of course, that's going to be the case.

What's being discussed with respect to DCA's measurements is a consistent, statistically significant bias.

 

Let me quote one of the NWS's climatologists on this:

"Snowfall data are riddled with issues like this. Everywhere, at some point or another, some observer has estimated instead of measured, or stuck their stick into a drift, or stuck their stick into a clearing. It is one of the many reasons that we don't put much stock in studies that investigate trends in climate based on snowfall data (except for the hope of offsetting penalties - that is, that the errors cause overestimation as often as they cause underestimation). Snowfall data are sausage - you definitely don't want to look too closely at what is in there, consume with caution, and make other things the staples of your diet."

[Ian]

Of course the ASOS was low.  But that doesn't mean its not comparable to IAD (which was also low).  The difference between the ASOS totals is comparable to the difference in snow totals.

Yes potentially. DCA prob had a bit more severe winds at peak just glancing? DCA's biggest winds coincided with the 2 hr or so big thump on the back end. Ratios (other than the wind, so at least rates) were great in that period.. I think that's where a lot of DCA's total is missing from what I know.

[famartin]

Yes, it is a big deal. First, we're talking about the potential of certifying a record that is wrong. I'm sure it happens, but that's not a reason to accept it. Second, it's the job of the measurers to get it right, and the job of the NWS to certify the right data. It doesn't matter if you are a tenth off or an inch off, you are given the responsibility to get it right, so move heaven and earth if necessary to get it right. 

 

 

This is an important point. Obviously you haven't had time to review all the evidence yet. I'm not quite ready to jump to a conclusion, as I know how these things go, but I still remain skeptical. All I ask is that the people involved in this process show their work. There is nothing wrong with that. 

 

The larger the amount of snow and the winder the storm, the harder it is to get a truly accurate snow total.  This storm was both windy and large, making it near impossible.  I would bet that the majority of totals we've received could be an inch or more different if someone else took the measurements.

[MidlothianWX]

I am currently conversing with the AKQ lead met and co-op program manager on Facebook. I have been told that some airports use FAA contractors for snowfall measurements (RIC, DCA) while others utilize paid NWS employees at nearby co-op stations (ORF). The program manager has informed me that all of these decisions are entirely in the hands of the FAA and that the NWS has no control over them.

 

What frustrates me is that this all seems to be entirely arbitrary. If some airports are able to utilize NWS co-op observer's totals why not all of them? The AKQ lead met even claimed that KRIC is a micro-climate. Why use totals from there then? I feel like I'm spinning tires in the mud with this one.

[famartin]

Yes potentially. DCA prob had a bit more severe winds at peak just glancing? DCA's biggest winds coincided with the 2 hr or so big thump on the back end. Ratios (other than the wind, so at least rates) were great in that period.. I think that's where a lot of DCA's total is missing from what I know.

IAD actually gusted stronger during some of our heaviest snow in the morning, over 50 mph.  It was pretty amazing actually. 

 

Anyway, while yes, the wind will reduce catch, it also tends to increase perceived rate... kinda like driving down the highway versus standing still.  So with the gusty winds, the snow was perhaps perceived heavier than it really was (at least, if it was coming straight down).  Visibility was still very low but more because of everything racing sideways (and perhaps blowing from the surface) versus the actual amount falling. 

[JC-CT]

Second, it's the job of the measurers to get it right, and the job of the NWS to certify the right data. It doesn't matter if you are a tenth off or an inch off, you are given the responsibility to get it right, so move heaven and earth if necessary to get it right. 

 

Some organizations might not consider that an efficient use of heaven and earth.

[Deck Pic]

One other thing:  Really, DCA records should probably not be compared to downtown.  Just like with virtually every other major city, offices move to airports and create a discontinuous record.  Best not to squint too hard at them.  PHL set a bunch of record lows in the 40s-60s which still stand.  Why?  Because it moved to the airport from downtown.

 

we know all about this.  We may not know much about the nuances of SSW events, but we are experts on DC Climo

[famartin]

I am currently conversing with the AKQ lead met and co-op program manager on Facebook. I have been told that some airports use FAA contractors for snowfall measurements (RIC, DCA) while others utilize paid NWS employees at nearby co-op stations (ORF). The program manager has informed me that all of these decisions are entirely in the hands of the FAA and that the NWS has no control over them.

 

What frustrates me is that this all seems to be entirely arbitrary. If some airports are able to utilize NWS co-op observer's totals why not all of them? The AKQ lead met even claimed that KRIC is a micro-climate. Why use totals from there then? I feel like I'm spinning tires in the mud with this one.

 

Several issues going on. FAA doing it makes it easier and cheaper, while if we hire someone else, then obviously we have to pay them and train them.  Plus, of course, its unlikely to get someone else who is as close to the ASOS (official measuring point) as the FAA observer will be.  Ideally we would want the measurement at the airport to be the most climatologically compatible, not a few miles away. 

 

We do have climate sites (not here at LWX but elsewhere in the country) where there simply are no longer FAA observers at the airport, and we have to hire people to measure snow. 

 

Measuring snow is a pain.  We didn't make it any easier on ourselves by putting in ASOS... really, doing that made it harder.  We're still fixing climate record snowfall gaps cause by ASOS installation across the nation in some spots. 

[Deck Pic]

I am currently conversing with the AKQ lead met and co-op program manager on Facebook. I have been told that some airports use FAA contractors for snowfall measurements (RIC, DCA) while others utilize paid NWS employees at nearby co-op stations (ORF). The program manager has informed me that all of these decisions are entirely in the hands of the FAA and that the NWS has no control over them.

 

What frustrates me is that this all seems to be entirely arbitrary. If some airports are able to utilize NWS co-op observer's totals why not all of them? The AKQ lead met even claimed that KRIC is a micro-climate. Why use totals from there then? I feel like I'm spinning tires in the mud with this one.

 

Richmond's issues have been going on for 20 years, as I am sure you know. I know at some point Anthony and others were able to convince AKQ for a certain storm to use Sandston spotter total instead of the airport.  Amendments to the record are not uncommon. 

[MidlothianWX]

Richmond's issues have been going on for 20 years, as I am sure you know. I know at some point Anthony and others were able to convince AKQ for a certain storm to use Sandston spotter total instead of the airport.  Amendments to the record are not uncommon. 

Anthony has been banging this drum for years. I admire his effort and have tried to assist him by calling AKQ, leaving comments on their FB page, sending emails, etc.

 

You are correct about the Sandston override. Brian Hurley has confirmed that this was the policy when he left AKQ a few years ago. The lead met has informed me that the totals were too close (11.4 vs. 12.4) to do anything this time around though.

 

This still begs the question: why does ORF have an NWS observer and RIC does not? I understand the points about cost savings etc. But why not have all airports in an NWSFO area doing the same thing?

[Deck Pic]

I say use the Crystal City total if it passes a QC. 1.5 miles away. Same elevation.  IAD doesn't measure at IAD.  PHL measures in Jersey.  No reason not to use the Crystal City total if it followed protocol.

[Ian]

IAD actually gusted stronger during some of our heaviest snow in the morning, over 50 mph. It was pretty amazing actually.

Anyway, while yes, the wind will reduce catch, it also tends to increase perceived rate... kinda like driving down the highway versus standing still. So with the gusty winds, the snow was perhaps perceived heavier than it really was (at least, if it was coming straight down). Visibility was still very low but more because of everything racing sideways (and perhaps blowing from the surface) versus the actual amount falling.

Fair enough tho we averaged prob 2"/hr in that period based on my own measurements. Admittedly difficult at that pt in storm but the city does offer some shelter.

I'm not too worked up about it. This was a top end storm for the region only comparable to a handful of others we know about it. Past that is academic or a pissing contest more often than not.

[LakeEffectKing]

Back a decade or so ago, I assisted Tom Nizol and Steve McLaughlin, both at NWS Buffalo at the time, with a "snowfall measuring" data collecting project to assist in developing algorithms to assess the variability of various snowfall reporting "swiping times" during several LES events.

Granted, during most of these events, I was dealing with ratios of between 15:1 to 25:1...but in summary: I had 9 boards in (3 groups of 3) throughout my sheltered yard (large yard)...in most of the events there was little to no wind...at each group of boards, I measured 1.)rates (swipe every hour), 2.) (3 hour total, then swipe) and 3.) The minimum time "official" totals can be reported (swipe every 6 hours). I averaged the three groups of boards for the three different measuring intervals. I observed some pretty impressive "settling rates"! I once had two consecutive averaged 3 hour totals of 8" and 10"...during those concurrent times, my 6hr. averaged swipe totals were 15"....thus an average settling rate of 1/2"/hr. over those 6 hours! And I have observed settling rates (with no wind under the center of a convergence LES band) of up to 2"/hr. during long duration, high total, high rate snowfall events!!

And if you go back through the Blizzard obs thread and watch a couple of the time lapse videos posted...during some of the lulls, you can see the snowpack settle....especially at the end of the videos when the snow stops...

The take home message is that snowfall measuring can be very sensitive to minor differences in methods used during: Heavy snowfall events, obviously windy events, and measuring start and stop times. But IMO, a really big problem is losing the "measuring equipment"....DOH!!! Just my two cents! Been measuring snow for over 40 years in LES country!

[MN Transplant]

Fair enough tho we averaged prob 2"/hr in that period based on my own measurements. Admittedly difficult at that pt in storm but the city does offer some shelter.

I'm not too worked up about it. This was a top end storm for the region only comparable to a handful of others we know about it. Past that is academic or a pissing contest more often than not.

I seriously doubt we were anywhere near 2"/hr Sat afternoon. We barely got that overnight while being death banded. The wind really screwed with people's perception of actual snowfall, I think.

[MN Transplant]

I'm still waiting for someone to explain to me why its a big controversy that DCA's total, which matches their depth, is a problem, but IAD's, which also matches their depth, is not.

The meteorological explanation is easy: DCA got dry slotted. The evidence besides the observer measurements is abundant: Nearby spotter reports in the range of DCA's total.

Lets take a look at ASOS precip totals, perhaps: DCA with 17.8" reported 1.48" liquid (raw ASOS, since observers don't melt it for us). IAD with 29.3" reported 2.23" liquid (that's raw ASOS, not what we melted, which was more).

The only reason this is an issue is because people forecast big totals and DCA didn't deliver for them.

Ray- what did you melt for liquid equivalent? I was close to 2.90".

[famartin]

Ray- what did you melt for liquid equivalent? I was close to 2.90".

2.80" at my house. Office was 2.63".  Even that could be a smidge low given our office rain gauge is pretty exposed to the wind.  Better than ASOS though.

[Deck Pic]

I seriously doubt we were anywhere near 2"/hr Sat afternoon. We barely got that overnight while being death banded. The wind really screwed with people's perception of actual snowfall, I think.

 Agreed, though I am pretty sure DC/East did better in the finale than the close in western burbs.  I got around 6.5 -7" after 7am on Saturday, and only around 1/1.5" fell during the lull.  After 2:30 or so, I pulled down another ~5/5.5". That's part of the problem with the DCA total.  Supposedly they only picked up 2.9" after 1pm.  Doesn't seem credible at all to me given the bands we got into and the reported totals.  Wes got 5.25" in the finale, and he mentioned how he wasn't in the good stuff for a while after DC got into it.  So DCA picked up 2.1" between 1pm and 4pm and another 0.8" between 4pm and 8pm?  way doubtful.

[SomeguyfromTakomaPark]

 Agreed, though I am pretty sure DC/East did better in the finale than the close in western burbs.  I got around 6.5 -7" after 7am on Saturday, and only around 1/1.5" fell during the lull.  After 2:30 or so, I pulled down another ~5/5.5". That's part of the problem with the DCA total.  Supposedly they only picked up 2.9" after 1pm.  Doesn't seem credible at all to me given the bands we got into and the reported totals.  Wes got 5.25" in the finale, and he mentioned how he wasn't in the good stuff for a while after DC got into it.  So DCA picked up 2.1" between 1pm and 4pm and another 0.8" between 4pm and 8pm?  way doubtful.

 

Yeah that doesn't make sense.  Like you said I picked up around 6 inches after 2 and I was in pretty similar bands to DCA the whole time.

[Ian]

I seriously doubt we were anywhere near 2"/hr Sat afternoon. We barely got that overnight while being death banded. The wind really screwed with people's perception of actual snowfall, I think.

Meant 1/hr, typo. Maybe not even that though it sure seemed quite heavy at times and I was out about 90 min straight with solid increases. Not even sure what DCA got in that timeframe since I was out but seemed discrepancies were in the late day period especially.

[gymengineer]

I seriously doubt we were anywhere near 2"/hr Sat afternoon. We barely got that overnight while being death banded. The wind really screwed with people's perception of actual snowfall, I think.

Wait, wasn't even DCA itself showing a 2" snow in one METAR and then 1" in the subsequent one (for between 3 and 5 pm)? 

 

Edited: Here-- that's pretty much what brought them from 14.9" at 1 pm to near 18" by 5 pm. It was the period after 5 pm that was sketchy with barely any more snow reported. 

 

METAR KDCA 232052Z 34024G34KT 1/4SM R01/1600V2000FT +SN BLSN FG VV007 M03/M05 A2962 RMK AO2 PRESRR SLP029 SNINCR 2/17 P0002 60005 T10331050 53021 $

 

METAR KDCA 232152Z 34023G32KT 1/16SM R01/1800V2000FT SN BLSN FG VV006 M03/M05 A2966 RMK AO2 PK WND 34036/2118 SLP044 SNINCR 1/18 P0002 T10331050 $

[famartin]

Ya know what... egg on my face.  There is a screwup.  Frig.  Dangit.  My bad guys.  My deepest apologies. 

[MacintoshPro]

Ya know what... egg on my face. There is a screwup. Frig. Dangit. My bad guys. My deepest apologies.

Could you elaborate..? Will the total be revised?

[famartin]

Could you elaborate..? Will the total be revised?

 

I would assume so, but you know what assuming does to you and me.  Its so obvious I'm dismayed I didn't notice it before.  I don't think anyone else has mentioned it because its glaring once you see it.  Kind of a "smack in the face durh" thing.  Don't want to reveal til its fixed, though.  I probably said way too much here anyway already.

[Stebo]

I noticed an issue today going over the observations as well beyond what I was monitoring with the obs on Saturday.

 

This is the 4:52PM and 5:52PM METARs with the SPECIs between:

 

METAR KDCA 232152Z 34023G32KT 1/16SM R01/1800V2000FT SN BLSN FG VV006 M03/M05 A2966 RMK AO2 PK WND 34036/2118 SLP044 SNINCR 1/18 P0002 T10331050 $                      

SPECI KDCA 232218Z 34025G37KT 1/8SM R01/1800V2400FT SN BLSN FG VV006 M03/M04 A2968 RMK AO2 PK WND 34037/2216 P0001 T10331044 $                            

SPECI KDCA 232239Z 34022G29KT 1/2SM R01/3000V4500FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 P0003 T10281044 $                            

METAR KDCA 232252Z 34023KT 1/2SM R01/3500V5000FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 SLP054 SNINCR 1/18 P0003 T10281044 $

 

First off the 1/16SM and 1/8SM obs should be +SN and all should be FZFG but the other issue is how do you gain an inch in an hour both hours but still remain 18" on the ground. This right here is a clear discrepancy in the observing. You can't lose and then gain an inch of snow within an hour even in a blizzard, it just doesn't happen. Based upon what I have read, my conclusion is the measuring they did was snow depth and not snowfall as snowfall would have been a higher total as you'd have compaction with depth consider the storm was over 24 hours and a heavier snow ratio at that.

[JonathanW]

Ya know what... egg on my face.  There is a screwup.  Frig.  Dangit.  My bad guys.  My deepest apologies. 

 

No apology necessary.

[wxnut]

I noticed an issue today going over the observations as well beyond what I was monitoring with the obs on Saturday.

 

This is the 4:52PM and 5:52PM METARs with the SPECIs between:

 

METAR KDCA 232152Z 34023G32KT 1/16SM R01/1800V2000FT SN BLSN FG VV006 M03/M05 A2966 RMK AO2 PK WND 34036/2118 SLP044 SNINCR 1/18 P0002 T10331050 $                      

SPECI KDCA 232218Z 34025G37KT 1/8SM R01/1800V2400FT SN BLSN FG VV006 M03/M04 A2968 RMK AO2 PK WND 34037/2216 P0001 T10331044 $                            

SPECI KDCA 232239Z 34022G29KT 1/2SM R01/3000V4500FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 P0003 T10281044 $                            

METAR KDCA 232252Z 34023KT 1/2SM R01/3500V5000FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 SLP054 SNINCR 1/18 P0003 T10281044 $

 

First off the 1/16SM and 1/8SM obs should be +SN and all should be FZFG but the other issue is how do you gain an inch in an hour both hours but still remain 18" on the ground. This right here is a clear discrepancy in the observing. You can't lose and then gain an inch of snow within an hour even in a blizzard, it just doesn't happen. Based upon what I have read, my conclusion is the measuring they did was snow depth and not snowfall as snowfall would have been a higher total as you'd have compaction with depth consider the storm was over 24 hours and a heavier snow ratio at that.

 

Note also that the AWOS recorded 0.08" of melted precip after 2152Z (when snow total was up to 17.5" I believe). Assuming 12:1 ratio, that would yield another 1" of snow after 2152Z (instead of 0.3") for a storm total of at least 18.5". Considering AWOS undercatches snowfall in windy conditions, you could argue that a storm total of 19" or so would be acceptable for DCA.  

[famartin]

I noticed an issue today going over the observations as well beyond what I was monitoring with the obs on Saturday.

 

This is the 4:52PM and 5:52PM METARs with the SPECIs between:

 

METAR KDCA 232152Z 34023G32KT 1/16SM R01/1800V2000FT SN BLSN FG VV006 M03/M05 A2966 RMK AO2 PK WND 34036/2118 SLP044 SNINCR 1/18 P0002 T10331050 $                      

SPECI KDCA 232218Z 34025G37KT 1/8SM R01/1800V2400FT SN BLSN FG VV006 M03/M04 A2968 RMK AO2 PK WND 34037/2216 P0001 T10331044 $                            

SPECI KDCA 232239Z 34022G29KT 1/2SM R01/3000V4500FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 P0003 T10281044 $                            

METAR KDCA 232252Z 34023KT 1/2SM R01/3500V5000FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 SLP054 SNINCR 1/18 P0003 T10281044 $

 

First off the 1/16SM and 1/8SM obs should be +SN and all should be FZFG but the other issue is how do you gain an inch in an hour both hours but still remain 18" on the ground. This right here is a clear discrepancy in the observing. You can't lose and then gain an inch of snow within an hour even in a blizzard, it just doesn't happen. Based upon what I have read, my conclusion is the measuring they did was snow depth and not snowfall as snowfall would have been a higher total as you'd have compaction with depth consider the storm was over 24 hours and a heavier snow ratio at that.

That's not the error I saw actually, its more glaring than that once you see it.  That said, I wonder if that's truly an error.  Since, from what I understand, FAA likes hourly totals (because of plow ops), its possible they were doing hourly swipes on another board?  We stopped using FAA at BWI (they still measure though) because they just did hourly swipes on one board and didn't bother with a 6-hour board.  Anyway, they could get 1 inch new on the hourly board, but because of settling, it might not add up as much on the other, so you might keep the same depth (within an inch at least) on the 6-hour board while gaining an inch on the other.  Anyway, those measurements are murky, maybe there was an issue, its uncertain.  What I noticed is *glaring*.

[Stebo]

That's not the error I saw actually, its more glaring than that once you see it.  That said, I wonder if that's truly an error.  Since, from what I understand, FAA likes hourly totals (because of plow ops), its possible they were doing hourly swipes on another board?  We stopped using FAA at BWI (they still measure though) because they just did hourly swipes on one board and didn't bother with a 6-hour board.  Anyway, they could get 1 inch new on the hourly board, but because of settling, it might not add up as much on the other, so you might keep the same depth (within an inch at least) on the 6-hour board while gaining an inch on the other.  Anyway, those measurements are murky, maybe there was an issue, its uncertain.  What I noticed is *glaring*.

I am actually not seeing the glaring error but to me I'd have a hard time thinking the snow settled an inch but also gained an inch within the hour in question, like you say it is an inconsistency in measuring hour to hour.

[wxdude64]

I noticed an issue today going over the observations as well beyond what I was monitoring with the obs on Saturday.

 

This is the 4:52PM and 5:52PM METARs with the SPECIs between:

 

METAR KDCA 232152Z 34023G32KT 1/16SM R01/1800V2000FT SN BLSN FG VV006 M03/M05 A2966 RMK AO2 PK WND 34036/2118 SLP044 SNINCR 1/18 P0002 T10331050 $                      

SPECI KDCA 232218Z 34025G37KT 1/8SM R01/1800V2400FT SN BLSN FG VV006 M03/M04 A2968 RMK AO2 PK WND 34037/2216 P0001 T10331044 $                            

SPECI KDCA 232239Z 34022G29KT 1/2SM R01/3000V4500FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 P0003 T10281044 $                            

METAR KDCA 232252Z 34023KT 1/2SM R01/3500V5000FT SN BLSN FG VV008 M03/M04 A2969 RMK AO2 PK WND 34037/2216 SLP054 SNINCR 1/18 P0003 T10281044 $

 

First off the 1/16SM and 1/8SM obs should be +SN and all should be FZFG but the other issue is how do you gain an inch in an hour both hours but still remain 18" on the ground. This right here is a clear discrepancy in the observing. You can't lose and then gain an inch of snow within an hour even in a blizzard, it just doesn't happen. Based upon what I have read, my conclusion is the measuring they did was snow depth and not snowfall as snowfall would have been a higher total as you'd have compaction with depth consider the storm was over 24 hours and a heavier snow ratio at that.

Yep, there is one inch anyway.