Thursday, August 30, 2007

Last Weekend's Storms Near San Diego

I have been away from the weather for most of the past week, as I had to prepare a book catalogue for mailing before the Labor Day weekend. I now have a bit of time to catch up on weather events and the like.

Jim Means, who is an avid storm watcher and monsoon follower, sent information at the first of the week, as per:

"Finally a big monsoon today over here in Southern California. The combination of an upper level disturbance offshore from San Diego and a deep moisture field (PW ~ 2") from the remnants of Hurricane Dean gave San Diego County its best monsoon day in quite a while. Lifting caused by the upper level disturbance initiated storms before sunrise, and the NWS in San Diego had to issue an unusual Flash Flood Warning for the area around Escondido, about 25 miles north of San Diego. A strong (but not severe) storm produced flash flooding and some areas got more than 2 inches of rain in about an hour and a half. Another storm occurring at the same time produced a waterspout offshore from San Clemente. Daytime heating later generated another flash flood producing storm in the area of Anza Borrego Desert State Park, where people had to be rescued from their cars in what are normally dry washes. This storm generated 3 inches of rain in some places and the NWS reported the top at 60,000 feet."

Jim, who lives in Alpine CA - which is east of San Diego along I-10, has sent a couple of photos of interest:

Photo 1, taken by Jim at sunrise on Sunday 26 August, shows the early morning storm near Escondido, north of Alpine.

Photo 2 shows Escondido Creek still running in flood on Monday morning.

Friday, August 24, 2007

More Information on Missing Winds in the NWS Upper-Air Soundings

I took a quick look at the soundings from yesterday evening and from this morning. At least the following soundings have missing wind data:

0000 UTC 24 August 2007

RAP sfc - 620 mb
LCH sfc - 400 mb
SHV sfc - 600 mb
TLH sfc - 720 mb
MHX sfc - 680 mb

1200 UTC 24 August 2007

BOI sfc - 600 mb
TFX sfc - 700 mb
MAF sfc wind is only wind for entire sounding
TLH sfc - 680 mb
MHX sfc - 780 mb
IAD sfc - 750 mb
BUF sfc - 780 mb

I browsed though the soundings at the Storm Prediction Center (SPC) site:

These soundings now display only the actually observed winds aloft.

I would imagine that such serious losses of wind data would be having negative impacts on the models and also upon forecast operations at places such as the SPC.

Discussion of the TWC Sounding for 0000 UTC 24 August 2007

The TWC sounding taken yesterday afternoon Figure 1 illustrates several of the serious problems with data from the NWS RRS Sippican sondes. I discuss the sounding layer by layer beginning at the surface:

* Abrupt jump in Td at release - reason for this is unknown. See my earlier post on the Madweather webpage regarding the roof-top launch conditions at TWC.

* The afternoon boundary layer (BL) in T extends upward to 700 mb; however, the Td trace from 900 to 780 mb does not depict a well-mixed BL in moisture. My guess is that this reflects a slow response of the hygrister back toward actual values, after the abrupt Td excursion at release.

* A realistic portion of the BL is is sampled from 780 to 700 mb.

* The sonde apparently entered a towering cumulus at about 660 mb and experienced severe wetting of the thermistor and perhaps of the hygrister. The data are extremely bad from 600 to 400 mb affected by evaporation, and perhaps freezing, and sublimation. Note - I was inthe immediate vicinity of the launch site and there were no thunderstorms present; however, a deep and thick mesoscale anvil covered the sky from the WSW through the ENE. The equally absurd ASOS surface observation from TUS at this time was "CLR."

* From 400 to 300 mb the sensors are apparently slowly recovering toward values present in the actual atmosphere.

* Above 300 mb the data remain suspect; are extremely smooth; and do not reflect the presence of the ice saturated anvil cloud.

* The temperature data from at least 600 to 300 mb appear too cool. Since the thermodynamic data are used to compute the heights of the pressure surfaces, it is not surprising that heights at and above 300 mb appear 20 to 40 m too low.

Thus, it is likely that this sounding had physically realistic data for a single layer that was only about 80 mb deep!

I certainly hope that these data do not end up in the NCDC sounding data archive.

Thursday, August 23, 2007

Winds Aloft Problem with RRS sondes - Today's Worst Example

While I didn't have time to plow through the details of all the NWS RRS Sippican soundings this morning, the worst example was not hard to find. The upper-air sounding at Boise, ID (WMO 72681), 1200 UTC, 23 August 2007, observed no winds above the surface.

I keep hoping that the NWS will formally alert external users of upper-air data, whether the users be proviate forecasters or the university and research community, of the problems that are plaguing their new RRS soundings.

As far as I know, such information has not been shared with the external user community - if my assessment is wrong, please let me know. Thanks. Bob

Sunday, August 19, 2007

Overview of Problems with Data from the RRS Sippican Sondes

I have seen a copy of an internal NWS issue paper on the RRS sonde, including "talking points" for the media. After reading this document, I have a number of concerns:

The sonde problem is apparently being approached by higher levels of the NWS as if it is essentially a problem related to the performance of the Sippican hygrister in dry, desert environments.

In my opinion, nothing could be further from the truth.

After looking at many RRS soundings during the past month and a half, I conclude that there are serious problems with both the hygrister and the thermistor and, probably, with the sonde's internal electronic components. This last statement is just a hypothesis, since I don't know much about the technical details.

I am collecting bad RRS soundings from all over the country so that I can provide an extensive documentation of the problems. I provide here a quick overview of theproblems I have seen to date:

1 - Extreme dry layers above the surface with a slow recovery, leading to soundings that are too dry. (Note that this "too dry in low levels" problem was first detected by NWS personnel in the Southwest. However, the NWS instruction manual for operators of the RRS system indicates that other of the sonde's problems were detected during evaluation; however, implementation proceeded regardless.) See Fig. 1.

2 - Slow response to dry layers aloft when sonde is launched in a relatively humid environment, leading to occasional soundings that are too wet. See Fig. 2 and post of August 8th.

3 - Both the thermistor and hygrister are apparently poorly vented and one or both are easily wetted, leading to a plethora of problems caused by the latent heat effects of condensation, evaporation, freezing, and sublimation. (The wetting problems apparently relate to the compact size of the new overall instrument package that was evidently developed initially for military field use.) See Fig. 3 and Fig. 4.

4 - The RRS soundings are sometimes missing winds through significant layers (the ultimate example of which I distributed the other day - for a MN sounding, hardly a desert environment!) - See a previous post made today. The problems the missing winds can cause are many, some of which have serious implications for forecasting. This problem may be related to the sonde's internal electronic components. See both Fig. 1 and Fig. 5.

5 - Other problems, again probably related to internal electronics, cause high to veryhigh frequency noise leading sometimes to many many "special" points - in bothT and Td. See Fig. 6.

Avoiding GEMPAK Interpolation when Processing Upper-Air Soundings

The following information on how to avoid interpolation within the GEMPAK code has been provided by David O. Blanchard - many thanks!

The fix for the interpolation of winds and temperatures in GEMPAK is not difficult.

GEMPAK can store upper-air data in two formats: merged and unmerged.

Unmerged data sets will store the TTAA, TTBB, and PPBB data as separate parts. There is no interpolation of the data.

Merged data sets combine the TTAA, TTBB, and PPBB parts and contain values for each parameter (P, T, Td, Dir, Spd, and Z) at each level and interpolates the data.

To avoid interpolation, the data must be saved unmerged. This is specified with the "MRGDAT" variable in GEMPAK when originally creating the upper air file.

Most sounding programs that use GEMPAK upper-air data will call SN_RDAT to read the data; SN_RDAT calls SN_MERG to merge the TTAA/TTBB/PPBB data if they are not already merged; SN_MERG calls MR_UADT which calls MR_MISS to interpolate the data.

Only MR_UADT needs to be modified. Copy the MR_UADT source code and the required INCLUDE files to your sounding software directory. Edit MR_UADT and comment out the line that calls MR_MISS. Compile your sounding source code and the modified MR_UADT source code.

The necessary code is found in the following locations:

./gempak/include/MCHPRM.PRM (rename MCHPRM.machineType to MCHPRM.PRM)

This patch prevents interpolation of the upper-air data in your sounding program but makes no changes to GEMPAK and its programs and libraries. Your sounding program must handle the uninterpolated data arrays
(i.e., it must be able to handle missing data flags of -9999.)

Example of merged TTAA/TTBB/PPBB data with interpolation:
603.0 3.8 -45.2 298.7 4.4 4375.2
589.0 2.8 -46.2 295.5 4.7 4565.5
558.0 -1.1 -36.1 288.1 5.4 4999.7
546.0 -0.3 -49.3 285.2 5.6 5173.3
545.4 -0.4 -49.4 285.0 5.7 5182.0
505.0 -5.1 -54.1 280.6 5.2 5791.9
500.0 -5.5 -53.5 280.0 5.2 5870.0
485.7 -6.2 -54.2 260.0 4.1 6096.0
477.0 -6.7 -54.7 261.5 4.6 6237.0
431.2 -12.6 -58.4 270.0 7.2 7010.0
416.0 -14.7 -59.7 252.0 6.3 7284.9
414.3 -14.8 -59.4 250.0 6.2 7315.0
400.0 -15.9 -56.9 255.0 7.7 7580.0

Example of merged data TTAA/TTBB/PPBB without interpolation:
603.0 3.8 -45.2 -9999.0 -9999.0 4375.2
589.0 2.8 -46.2 -9999.0 -9999.0 4565.5
558.0 -1.1 -36.1 -9999.0 -9999.0 4999.7
546.0 -0.3 -49.3 -9999.0 -9999.0 5173.3
545.4 -9999.0 -9999.0 285.0 5.7 5182.0
505.0 -5.1 -54.1 -9999.0 -9999.0 5791.9
500.0 -5.5 -53.5 280.0 5.2 5870.0
485.7 -9999.0 -9999 .0 260.0 4.1 6096.0
477.0 -6.7 -54.7 -9999.0 -9999.0 6237.0
431.2 -9999.0 -9999.0 270.0 7.2 7010.0
416.0 -14.7 -59.7 -9999.0 -9999.0 7284.9
414.3 -9999.0 -9999.0 250.0 6.2 7315.0
400.0 -15.9 -56.9 255.0 7.7 7580.0

Missing Winds from RRS Soundings - An Extreme Example

On Monday afternoon, August 13, 2007, Rob Dale and Eric Hanson alerted me to the strange characteristics of the special 1800 UTC sounding from MPX (Chanhassen, MN, an RRS/Sippican sonde site).

The unbelievably smooth hodograph is what had caught their attention. The sounding, as analyzed by software analysis programs at the Storm Prediction Center (SPC) is shown in Fig. 1. Note the perfectly smooth hodograph and the winds that back continuously from the surface to 150 mb. This seems a very strange sounding indeed!

However, examination of the sounding at the FSL RAOB site indicated what the problem was. The MPX sounding at FSL (Fig. 2) indicates that the Sippican sonde apparently lost GPS tracking immediately off the surface and did not recover the signals until 150 mb - i.e., the winds were actually missing from the surface to 150 mb.

So where did the "data" plotted at SPC come from? It turns out that many sounding analysis programs use, as part of their code, GEMPAK. The GEMPAK procedure interpolates data into layers where observations are missing - assuming probably that such layers are never very deep.

The sounding plot from the Univ. of Wyoming site (Fig. 3) shows interpolated winds that are similar to those on the SPC plot. However, Lance Bosart alerted me to look at the same sounding at the UCAR/RAP upper-air site (Fig. 4). Their plot shows a completely different wind profile through the troposphere! Apparently, RAP software uses a different interpolation procedure. The winds in their "answer" for the deep layer of missing observations essentially keep the direction at 150 mb constant, while decreasing the speed smoothly until near the surface, where a shallow layer of backing winds is shown.

While I'm not sure what impact this comedy of errors might have had on forecasts during the afternoon of August 13th, it does vividly illustrate how serious the Sippican missing winds problem can be. I have been alerted by a number of people around the country about the missing winds problem, and I will document it further at a later time. However, this is a particularly insidious problem for non-NWS users of the data. If one suspects that a sounding may have missing winds, then a time-consuming examination of the sounding at different sites has to be made for accurate evaluation.

Wednesday, August 15, 2007

Severe Storm Complexes Two Nights in a Row

I have chased down some photos of damage that occurred in Casa Grande on Monday night. See Photos 1-3.

As for yesterday afternoon (Tuesday Aug. 14) - I was out on east side when storms rolled through around 3:30 to 4 pm. Thought I saw several wall clouds, and there was a spectacular hail shaft out over Oro Valley around 4:30 pm. Of course my camera was back here at house! I see that there was wind damage reported in Oro Valley with the afternoon storm.

I note on the SPC page that there was severe hail NW of Flagstaff yesterdayafternoon and also a damaging microburst in Riverside CA!

I was very surprised by the early storm developments yesterday afternoon. Obviously, I did not properly adjust and then forecast the morning sounding - at least there's now another "out" available, given the bad Sippican sounding data!

MCSs developed again during the night (Tuesday night) and the activity continues this morning in south-central AZ. Tucson, Casa Grande, and Phoenix Skyharbor all reported gusts over 50 mph near or after midnight. Phoenix reported a severe gust to 59 mph,and there was wind damage near the Tucson airport.

The question for today is whether we'll experience similar intense and severe storms for the third afternoon/night in a row? The soundings this morning appear similar to yesterday (TWC again too wet wrt GPS IPW) with at least moderate CAPE available given decent heating, and the low-levels remain cooled by the recent outflows. The steering winds remain relatively strong and there is better directional shear aloft with another inverted trough approaching. I am again thinking that activity should be late and after dark, but the NAM seems to convect early and to move another organized line across southern AZ during the mid-afternoon. We will see, and I'll keep my camera at hand today.

The weekend could be interesting if part of the tropical system over Gulf of Mexico breaks off and comes across Sonora and southern AZ.

Sunday, August 12, 2007

Additional Commentaries on the Damage from Friday's Severe Thunderstorm

From Jim Toth:

Driving west on Orange Grove, I did not see anything significant until approaching the intersection at Oracle. The Park Place apartments are on the north side of OG, and that's where you start to see large trees down.

North on Oracle, the damage sticks out on both sides of the road all the way north to Ina. Aside from trees down and stripped green leaves obscuring parking lots, the most impressive thing I saw was the Long Realty building on the west side of Oracle. They had a large number of clay roof tiles peeled back and settled at odd angles.

At the Safeway parking lot at Oracle and Ina, there were a few large mesquite trunks snapped. These were out near Oracle. West on Ina, at the UMC medical offices, just east of Safeway, there were two large trees down near the entrance. Farther east on Ina, and then south on First, I did not see anything.

Steve Mullen reporting from Boulder:

My oldest son Ryan got caught in the teeth of the macroburst that nailed the NW side yesterday. He was driving along Ina between La Cholla and Oracle before heading southbound Oracle to Orange Grove around 6:10 to 6:15. Ryan reported two snapped stop signs, a leveled ~100-year saguaro, too many downed big trees to keep track. The sign in front of HiFalutin Restraurant was completely snapped. It was raining and blowing so hard near Ina and La Canada that he had to pull off the road.

Ryan shortly thereafter experienced some small hail. The Tucson Citizen had a blurb stating that winds were 60 mph over the NW side. I agree that they were 60 mph, for a moment, before gusting faster. It would take far more than 60 mph to down all that got leveled last evening.

Saturday, August 11, 2007

Forecasting SE AZ Convection using Unreliable Upper-Air Sounding Data from NWS/TWC

The last two soundings taken at TWC illustrate how frustrating it can be trying to use unreliable data from the new NWS Sippican sondes. The sounding from 00 UTC on 11 August is shown in Fig. 1, and the 12 UTC sounding on 11 August is shown in Fig.2.

The FSL time series comparing GPS IPW with TWC sounding IPW is shown in Fig. 3. The sounding last evening was too dry by 5 mm wrt GPS IPW and this morning's sounding is too wet by 5 mm wrt GPS IPW. The forecaster apparently didn't notice this problem this morning:


RE: Fig. 1

The afternoon sounding yesterday is particularly difficult to resolve meteorologically.
First, the surface super-adiabatic contact layer seems to extend upward to 900 mb. Possible I guess, but it also may be because a dry-spike data point was eliminated by the person who processed the sounding. I have no way to determine whether this was the case.

Second, where is the missing moisture located in the vertical? Since the well-mixed BL in theta extends up to 700 mb and the well-mixed q BL extends only to about 780 mb, I suspect that the "missing" moisture was likely in the upper-part of the BL. This sounding was taken in proximity, in both time and space, to a severe thunderstorm, I think that the CAPE was probably greater than the 314 indicated by the sounding analysis software. These sounding problems may have affected the timeliness of the severe thunderstorm warning that was issued at 6:22 pm.

Obviously, unless this sounding is somehow flagged as suspect in the data archives, a future researcher could consider it a valid severe thunderstorm proximity sounding, leading to………..

RE: Fig 2

This morning's sounding is also very difficult to analyze wrt to the likelihood of afternoon storms. Question is where is the "extra" moisture hiding in this sounding?
It is probably not in the spikey (noisy) data above 620 mb (some Sippican soundings are characterized by extremely noisy T and q data points for unknown reasons). The strange layer from 900 to about 840 mb is my guess for the layer that is too moist. I estimate that a BL well-mixed in q this afternoon would be one with q of ~ 11 to 12 g/kg.

While the mountains will have storms today, it is not clear how the lower elevation BL may evolve. It will likely have CIN above it's top, and if warming occurs in mid-levels, may end up not supportive for storms trying to move to lower elevations. But this is all guess work - since I don't know what the thermodynamic structure is below 700 mb to begin with, it is very difficult to try to predict what the afternoon structures are likely to be.

As for my forecast for our house - that's much more simple. Regardless of what happens to the south, unidirectional southerly flow is very unfavorable for storms and measurable rain at this particular spot. So, I would say that POPs here at our house are near zero today.

Storms in Tucson Metro Area on Friday 10 August 2007

Storms did develop over much of southeastern and south central Arizona yesterday. However, Cochise County in the far southeast was relatively suppressed, perhaps because dry air from New Mexico and northern Mexico invaded that part of the state.

Here at our house, we began hearing thunder from storms on the Catalina Mountains by 4:30 pm. I took Photo 1 of the initial developments over the Catalinas at 3:45 pm, looking NNE. My estimate is that cloud bases were at 700 mb or a bit lower, which fits with what one would expect from yesterday morning's TWC sounding.

The evening sounding was quite dry and this morning's (Saturday 11 August) is too moist. More on the problems of trying to forecast storms using the flawed data from the NWS/Sippican sonde in the next post.

I took Photo 2 at about 5:15 pm, again looking NNE. At this time frequent lightning was visible, along with lots of rumbling thunder. Updrafts had built southward from the mountains. There was another cell off to the ESE that had come off the Rincon Mountains, and new updrafts were forming between them. Photo 3 shows the Catalina storms in a view to the N, i.e., a bit to west of my Photo 2. The storms from the Rincons and the Catalinas apparently merged over the north side of town and then moved westward as a severe thunderstorm, producing damaging downburst winds.

Here at house we experienced strong wind gusts of 40 to 55 mph (estimated) from about 5:50 pm until about 6:20 pm. Winds were initially from the east with lots of dust, then from the northeast, and finally from the north (I think, since visibility was near zero in heavy rain at this time). Rain accumulation here was only 0.29". Jim Toth reported 0.97" a couple of miles to our NW.

Photo 4 shows the view of the storm from campus at 5:54 pm looking N. The ragged appendage, ahead of the outflow, showed some rotation as I observed it from the house. Note that the updraft and leading edge of rain shaft show green coloring, and small hail was indeed reported with the storm. Photo 5 shows some of the wind damage produced by the storms outflow, several miles to our NW.

All in all, quite an interesting afternoon here!

Friday, August 10, 2007

Interesting Storm Day Developing over Southern Arizona

Looks like a very interesting day setting up locally.

Very nice push of low-level moisture into most of southern Arizona after midnight, makes today quite different than yesterday thermodynamically.

The TWC morning sounding exhibits a low-level structure that's often associated with such nighttime moisture pushes, and I think that it is a reasonably good representation of morning conditions - one of the few times this week. The sounding, with the middle-level cooling and added moisture, indicates moderate CAPE lower elevations and even more unstable conditions higher elevations. The strong storms to our south and southeast last night crashed as they moved toward lower elevations. Today should be better, with storms coming into the lower elevations.

The morning NAM indicates a good setting by evening with upper-level difluence, and 15 to 20 kt steering winds at 500 mb. All of this due to an upper-level inverted trof brushing over the top of the malingering weak wave at lower levels. The dilemma of course is that the NAM has been forecasting stronger middle level winds than have been observed the past few days - so we have to wonder whether conditions will be as good this evening as NAM suggests.

The VAD winds this morning show L/V winds aloft - except for KYUX which shows that the surge extends up to at least 6000 ft agl. So, the evolution of the wind field today will likely determine how aggressively storms move into lower elevations. A good day to watch the Doppler winds and the Ft. Huachuca profiler winds closely.

My guess is that today will be quite active SE third of state with possibility of strong to severe hybrid downbursts and locally heavy rains. Storms here at house? Probably at least a 50% chance or better.

Wednesday, August 08, 2007

An Examination of Suspect Thermodynamic Structures in the Tucson Upper-Air Sounding AT 12 UTC 8 August 2007

Those who have read my previous posts are aware that I have been documenting data problems in upper-air soundings taken by the NWS using the new Sippican sonde. This morning's TWC sounding illustrates a suspect thermodynamic structure that I have observed several times here during the last few weeks.

First, a brief discourse on boundary layer (BL) structures. The afternoon BL is usually, but not always, well-mixed in both potential temperature (theta) and in moisture mixing ratio (q). Such BL layer structure is illustrated in the Phoenix 00 UTC sounding from last evening (Fig. 1) and in the TWC 00 UTC sounding (Fig. 2), ignoring the dry spike in contact layer. TWC uses the new Sippican sonde, whereas, Phoenix uses a Vaisala sonde with a different type hygristor. Both afternoon soundings depict a well-mixed BL extending from just above the surface to about 750 mb.

When winds are light and there is little advection or turbulent mixing during the night (the situation present over southeastern AZ during the past night), a cool and stable nocturnal BL develops near the ground, with the residual BL from the previous afternoon remaining in place above the nocturnal BL. The Phoenix 12 UTC sounding this morning (Fig. 3) indicates a residual BL present from about 850 to 700 mb.

However, the TWC 12 UTC sounding this morning (Fig. 4) indicates a much different structure aloft over Tucson this morning. Note that the theta profile indicates a residual BL aloft in about the same layer as found in the Phoenix sounding. However, the structure of q does not indicate a residual BL in the moisture field. Above the nocturnal BL q decreases steadily from about 12 g/kg to about 8.5 g/kg at 750 mb and then jumps up to 10 g/kg at 700 mb. This structure is physically very suspect, although I can not prove that the q values are bad, I feel strongly that they are.

Based upon the TWC afternoon sounding at 00 UTC, one would expect to find a residual BL well-mixed in moisture, as in the Phoenix sounding, with nearly constant q values of about 10 g/kg. The Sippican sonde seems to have responded slowly to the abrupt change in q above the nocturnal BL decreasing steadily from 900 to 750 mb and then jumping back up to the expected value of ~ 10 g/kg at 700 mb.

Interestingly, although I am personally convinced that the TWC morning moisture profile data points are erroneous from 900 to 700 mb, the sounding's total IPW agrees very well with the GPS IPW. Thus, at times the GPS-derived data may indicate that a Sippican sounding is accurate wrt IPW, even though there may be serious problems within the details of the Sippican data.

Errors such as the ones likely present in this morning's TWC sounding can have serious implications if one is trying to forecast the likelihood of strong convective storms.

Weather Update for Next Couple of Days

The last week has been very interesting with very high IPW combined with very warm middle level temperatures. It has seemed like Central America. Several very heavy rain events occurred with little or no cloud to ground lightning. There were three fatalities here in the Tucson area due to flash flooding.

Yesterday was quite interesting as skies cleared and IPW decreased steadily. Winds aloft here in southeast AZ were very light. With the abundant sunshine, a well-mixed afternoon BL built up to about 750 mb. So, the drying was apparently a result of weak advection plus subsidence - particularly from 700 to 400 mb. The BL has retained fairly impressive moisture, and dewpoints remain high in the SE part of state.

The TWC sounding has suspect structures this morning - more on this in a separate blog post later this morning - but it appears to me that with heating and slow deepening of BL, coupled with at least some degree of cool advection in middle levels, deep convection will return very quickly. This afternoon we may see Cbs over the nearby mountains. Then things get more complicated.

There is a substantial tropical wave moving westward into the eastern Pacific. Very impressive convective complexes occurred over the lower GoC and further south last night. Thus, with falling pressures in lower Colorado River Basin and southern AZ (about 2 mb during last 24 hours), southerly winds, and moisture advection, should increase up the GoC.

The NAM has had great difficulty last two weeks over Mexico. Disturbances persist in analyses and forecasts with apparently a lot of latent-heating induced spinup leading to stronger middle level winds in the model than have likely been present in the real atmosphere. The lack of observations over most of Mexico lets the model feedback on itself and develop structures that may or not be real. For example, this morning's run takes the main tropical wave to the west-northwest, but it also breaks off a piece to the north up the GoC - winds around this feature may again be too strong, but it is certainly an interesting situation worthy of careful monitoring.

Meanwhile, I'm really enjoying the return of some sunshine to the desert!