THIS WEEK’S WEATHER

Thursday Thunderstorms

Posted Thursday 06/26/25 @ 5:13 PM — The NWS has issued a severe thunderstorm watch. Some strong thunderstorms have developed far northwest of Philadelphia. Latest radar at 5 PM—

Radar with RAP model superimposed vertical motion (Omega). Note the upward vertical motion is centered west of the city. (Click on image for a larger view.)

The models have only been so-so with today’s forecast. Many models have forecast areas of storms that have not materialized late this afternoon. That’s particularly true of the experimental RRFS which forecast many areas of rain at this very time.

The trend mentioned in this morning’s update suggests the heaviest activity will remain just west of Philadelphia, as shown in the NBM graphic earlier.

The latest HRDPS and NAM-NEST capture the most likely forecast, again keeping the heavy rain to our west.

12z HRDPS forecast for 8 PM. The main activity remains to our west and eastern parts of the city may not see all that much rainfall. (Click on image for a larger view.)

These things are tough to forecast with high precision. I guess we’ll see what happens in the next few hours.


Thursday Forecast

Posted Thursday 06/26/25 @ 9:44 AM — Nailing down the forecast locations and timing of thunderstorms yesterday met with mixed results; the additional storms forecast towards midnight didn’t materialize.

Today, Thursday, presents another thunderstorm forecast challenge. Model forecasts range from storms developing as early as 2 PM near the city to storms mostly staying to our west and south (and South Jersey) through most of the afternoon and evening.

The NBM total rain accumulation captures the overall trend of any showers and storms staying to our west, south and east—

The 12z NBM just available keeps much of the Philadelphia area rain-free through 8 PM with any showers/storms staying to our west, south and South Jersey. (Click on image for a larger view.)

In contrast to this NBM forecast is the latest ECMWF-AI model, which does have showers and storms in the immediate Philadelphia area by mid afternoon.

Also in contrast to the NBM is the Canadian GDPS- with “AI spectral nudging” which has thundershowers in and near Philadelphia between 3 PM and 7 PM.

00z GDPS with AI spectral nudging. 1 Hour Convective precipitation at 6PM (Click on image for a larger view.)

A tough call. I guess we’ll find out.


Wednesday Night into Thursday

Posted Wednesday 06/25/25 @ 8:16 PM — Isolated to scattered storms developed. Very low wind shear has resulted in the storms forming and then collapsing. (Wind shear is needed to keep thunderstorms alive.)

Current radar at 8:10 PM

Radar at 8:10 PM with RAP model vertical motion (Omega- green contours.) In addition to having no wind shear, these storms are moving into areas of negative Omega (downward vertical motion- white boxes around negative green dashed contours.) The storms should collapse, but not before some areas get some quick downpours. (Click on image for a larger view.)

Today’s HRRR was unusually poor in forecasting the current storms. The experimental RRFS did better. Continuing with the RRFS, it is shows additional scattered short-lived storms through 3 AM as the backdoor cold front slowly moves through.

Thursday will have a mix of clouds and sun. Isolated to scattered storms are expected to begin to pop about 3 PM into Thursday evening. More details tomorrow.


Wednesday into Early Thursday

Posted Wednesday 06/25/25 @ 8:30 AM — As described yesterday, the heat dome will retreat late Wednesday as cooler high pressure slowly noses in from the northeast (backdoor cold front). Before that time, temperatures will again approach high levels, between 98º to 100º in and near the city.

Dew points will be higher today at peak temperature time, and I’ll have to resort to the Heat Index temperatures to capture the full impact. (I avoid posting heat index values since I think they’re often used to unnecessarily sensationalize, exaggerate heat episodes on TV broadcasts. Maybe not so today.)

06z NBM “Apparent Temperatures”, aka Heat Index. (Click on image for a larger view.)

As for thunderstorms, high heat, high humidity and several short waves ahead of the approaching backdoor cold front will create conditions ripe for thunderstorms with heavy rain. The latest HRRR is least on-board with storms here, but the AI models, the NAM-NEST, Canadian RGEM,and the RRFS show the possibility of isolated storms (especially in the wind shift/convergence zone in NJ) as early as 2PM.

Increasingly thunderstorm coverage expected late in the afternoon and especially in the evening hours (6 PM to midnight) where they should become more numerous. Not everyone will see a storm, but those that do will likely see some heavy rain. Severe winds, always a possibility, are not expected at this time.

To convey the scattered nature of the storms, here’s another model parameter that attempts to predict the strength of individual thunderstorm cells at 7 PM.

06z NBM Estimate “Radar Echo Heights” at 7 PM. Anything over 30,000 feet can be a very strong thunderstorm cell. Here, there are some above 40,000 feet. The exact predicted location of these storms cannot be taken literally with these models. (Click on image for a larger view.)

Wednesday- Heat Early Afternoon Reduced by Thunderstorms

Posted Tuesday 06/24/25 @ 7:57 PM — Disturbances moving across the upper heat dome are expected to bring showers and thunderstorms to our area Wednesday afternoon and evening. Some may be strong.

Water Vapor image Tuesday evening shows upper air disturbances poised to move towards us on Wednesday afternoon and evening as the heat dome retreats somewhat allowing them to reach us. (Click on image for a larger view.)

High temperatures of 99º and possibly 100º in parts of the city will be reached early, before 2 PM.

Thunderstorms develop and move in from the northwest and develop in place as early as 2 PM and continue into the evening. Impressive thermal instability (negative 8 “lifted index” and near 40 K index) and high CAPE values approaching 3000 J/Kg could make any storm strong to severe.

I’ll update Wednesday morning.


Tue 4:30 PM —Forecast Review — Another hot day, where the AI models were more accurate with their temperature forecasts than the very advanced NBM statistical Model Blend. (Although the Model Blend did fine when you factor in the standard deviation.)

Here’s the RTMA temperatures at 3:45 PM—

RTMA temperatures at 3:45 PM today. Black contours are 1º increments (Click on image for a larger view.)

Here’s the official readings from the NWS Philadelphia Airport KPHL

Highest Heat Index occurred at about 2 PM with the dew points being the highest. (Click on image for a larger view.)

Heat Continues

Posted Tuesday 06/24/25 @ 9:26 AM — Many models have similar high temperature forecast for today. The latest Model Blend (NBM) just became available. Here’s its forecast highs with degree of uncertainty (standard deviation) for today, Tuesday

12z NBM High Temperatures (not heat index or apparent temps) (Click on image for a larger view.)

Here’s the latest ECMWF-AI model high temperature forecast—

06z ECMWF-AI High Temp forecast (Click on image for a larger view.)

Here’s the new Canadian CMC GDPS-AI forecast—

00z CMC GDPS-AI model high temps for Tuesday (Click on image for a larger view.)

Heat Recap

It really felt hot today! Temperatures seem to have only reached 99º in this area, but dew points were 70º (and where highs were only 96º, e.g. Blue Bell, the dew points were closer to 74º and 75º

Here’s the RTMA @ 3:45 PM, capturing the area’s temperatures—

RTMA 15 minute repeat model. Actual with interpolated temperatures. (Click on image for a larger view.)

The NWS has modernized their website for KPHL airport, now showing temperatures every 5 minutes—

(Click on image for a larger view.)

Of interest is the localized area of showers that developed, as predicted by the new Canadian GDPS-AI along with the wind convergence zone—

Radar at 4:35 PM Wind convergence liine with showers in southern Cape May county, as predicted by the Canadian AI. (Click on image for a larger view.)

It’s interesting that the experimental RRFS did not predict these showers in Cape May. (The NAM-NEST did predict these as well.)

For Tuesday, here’s the latest ECMWF-AI temperature forecast—

ECMWF-AI Single Temperature Forecast for Tuesday(Click on image for a larger view.)

Stay cool!


Originally Posted Mon 9:22 AM —Yesterday’s high temperatures were diminished by the unexpected and unforecast clouds and showers.

The heat will be here for sure today.

Continuing with the “experiment” from yesterday, here is the new Canadian GDPS-AI along with the Model Blend (NBM) forecast high temps for today. likely about 3-4 PM—

00z Canadian GDPS-AI high temps for Monday (Click on image for a larger view.)

Here’s the Model Blend (NBM) high temps—

06z NBM forecast high temps. Contours are 1º increments. (Click on image for a larger view.)

The “Open Data” version of the ECMWF-AI single is very similar to the above.

Just to see if the Canadian has any unique insights, it is forecasting some thundershowers to develop in NJ at 3-4 PM due to the development of a moisture convergence zone—

CMC-GDPS Moisture convergence zone thundershowers.

Stay cool!


WEEKEND WEATHER FORECAST

Sun 1:05 PM —Forecast ReviewNot a forecast to be proud of— Both last night’s and even this morning’s models failed to forecast this “mesoscale complex” with strong upper air support dropping south along the edge of the mid level heat dome.

I was out on a hike this morning and looked at the darkening skies to our north. I downloaded new data to my iPhone remotely and and saw the radar in no way matched the model forecasts from even two hours earlier. I remotely uploaded the “Update” below at 10:43 AM.

Here’s the current radar with the hourly RAP model superimposed—

Current satellite water vapor image with superimposed RAP model 500-1000 mb thickness lines (yellow), Mean Sea Level Pressure isobars (black contours) potential vorticity (fine violet contours) with superimposed MRMS RADAR. Water Vapor Imagery reveals emitted water vapor infrared TEMPERATURE, NOT REFLECTED LIGHT from CLOUDS. Orange contours are jet stream level vertical motion. Strong UPward areas right over Philadelphia suburbs into NJ will enhance showers/storms. The blue shading is the strong upper air vertical motion of the Mesoscale Complex.
  (Click on image for a larger view.)

This morning’s HRRR (12z -8AM EDT) shows the a diminished mesoscale complex moving far north of us at 1 PM—

12z HRRR forecast for 1 PM Sunday. Rain stays well to our north. Completely under forecast the rain and position. (Click on image for a larger view.)

The HRRR is one of our best “CAM” models. (CAM = “Convective Allowing Model”, the sort of process occurring that this model is designed to capture. ) Clearly, there was some input upper air data problem at the very least.

Our newest CAM model, the experimental RRFS showed a bit of activity with a 14z (10 AM) model run. Still quite disappointing.

14z (10 AM EDT) hourly RRFS (experimental) did somewhat better, but still not a good forecast. (Click on image for a larger view.)

Even this morning’s ECMW-AI model, just available, didn’t get this right—

This morning’s 12z ECMWF-AI model kept the rain far north of our area. (Click on image for a larger view.)

As I’ve said as recently as last week, these modeling errors and forecast ‘surprises’ is one of the things that I find fascinating about meteorology.

Update Sun 6/22 10:43 AM — The large area of rain to our far north was forecast by the models to be further north than currently on radar. They were also forecast to dissipate. They are the result of hot humid air pushing northward in the mid levels of the atmosphere.

The latest HRRR still shows it dissipating as it approaches.


Posted Sunday 06/22/25 @ 9:14 AM — The Canadian Meteorological Center (CMC) announced their release this week of their advance into AI weather models with the release of an experimental global model based on the GDPS. (GDPS-AI) . The CMC describes this new model as “spectrally nudged” towards an AI model. (Weather models use wave-form physics (spectral methods) to recreate the atmosphere structure and dynamics.)

The GDPS-AI is run twice daily, 00z and 12z and becomes available four hours and 45 minutes after it is run. It forecasts out 240 hours. The nice thing about this AI model: it has hourly forecasts through forecast hour 84, in contrast to the European and US AI models that forecast at 6 hour intervals.


My experience with the Canadian models has found their regular models tend to bias towards the high side of temperatures in this region during the summer months.

So let’s do an ‘experiment’. Let’s see if their AI version is forecasting temperatures that are too high today?

00z GDPS AI model temperature forecast for 5 PM today, Sunday. (Click on image for a larger view.)

Here’s the latest NBM model, just available. Its highs are significantly less, but the temperatures have a wide uncertainty (standard deviation) of ± 2.2º (The NBM (“model blend”) is a weighted average of 30-40 models)

Today’s 12z NBM high temp forecast. Contour lines are 1 degree increments. The Temps are statistical means, ±2.2º (Click on image for a larger view.)

It wouldn’t surprise me if the NBM is more accurate today, assuming adding 1º-2º to to its forecast.


Originally Posted Sat 10:28 AM —The main weather feature of this weekend and much of next week will be a building upper level ridge, bringing very high temperatures and humidity to the Philadelphia region. The building ridge is visible on this morning’s water vapor image—

Current satellite water vapor image with superimposed RAP model 500-1000 mb thickness lines (yellow), Mean Sea Level Pressure isobars (black contours) potential vorticity (fine violet contours) with superimposed MRMS RADAR. Water Vapor Imagery reveals emitted water vapor infrared based on TEMPERATURE, NOT REFLECTED LIGHT from CLOUDS.
As a result, Water Vapor Temperatures indirectly reveal height, called the “effective layer” of the water vapor in a region, therefore revealing the mid and upper level 3D contours of systems.. Colder temperatures (blue- grey – white) indicate higher altitude, sometimes revealing upward motion. Red- orange indicates warmer temperatures and a lower height for this layer. There are several water vapor “channels”; the one I display here is called the mid level height channel. (Click on image for a larger view.)

The expected temperatures based on the European AI model (the ECMWF-AI Single) were posted yesterday. When the temperatures here are hitting 100º, I don’t see that much value, (and only hype) in giving the heat indices.

According to the latest NBM, high temps will hit 90º today, Saturday near the city and 97º on Sunday.

When the temperatures get this hot, there’s always the possibility of an isolated thundershower popping up. A trigger is often needed for these to occur. The models show some storms in far northeastern PA later this afternoon.

Yet with the upper atmosphere being so hot, the upper level heat in many ways inhibits convective rainfall.


THIS WEEK’S WEATHER

Weekend Outlook and Beyond

Posted Friday 06/20/25 @ 8:29 AM — Increasing temperatures and high humidity will be the big story for the next six days. The European AI model has been very good with high temperature forecasts and here’s a sample for Sunday, Monday and Tuesday—

ECMWF-AI Temperatures at 2 PM. Contour lines are 1º increments Actual highs, which occur at 3-4 PM may be higher. (Click on image for a larger view.)
ECMWF-AI Temperatures at 2 PM. Contour lines are 1º increments Actual highs, which occur at 3-4 PM may be higher. (Click on image for a larger view.)
ECMWF-AI Temperatures at 2 PM. Contour lines are 1º increments Actual highs, which occur at 3-4 PM may be higher. (Click on image for a larger view.)

These are actual forecast highs at 2 PM. Since the ECMWF-AI data that I access doesn’t include temperatures at 3-4 PM, the actual likely daily high may be even higher. Since the TV people like to sensationalize with the heat indices, add about 5-10º to get the “apparent temperature” or heat index. Frankly, I don’t know the difference between 100º and 110º; neither is too comfortable.


Fri 8:24 AM —Forecast Review —High wind gusts were the main “feature” of Thursday evening’s storms, resulting in power failures. I watched as a the top of tree split off at the far end of our backyard. Rainfall amounts were only in the moderate range, as these storms were fast movers.

Here’s the MRMS precipitation estimate—

MRMS rain-gauge measured rainfall combined with rain-gauge calibrated and interpolated radar-based summary of ACTUAL RAINFALL RECEIVED. Color shading is in inches. Line contour numbers are in mm. (25.4 mm= 1 inch)      (Click on image for a larger view.)

Thursday’s Storms

Posted Thursday 06/19/25 @ 5:04 PM — The storms are moving in on schedule. There was a radar signal for hail in Lancaster county with these storms, but not present at the current moment.

Radar and satellite water vapor with RAP model vertical motion at 850 and 250 mb superimposed. The yellow contours show plenty of upward motion that was missing in yesterday’s storms. (Click on image for a larger view.)

They are fast moving and there may be additional showers later this evening.

Posted Thursday 06/19/25 @ 1:33 PM — A quick update. Additional models show likely severe weather that includes the immediate Philadelphia area, but still more likely just south, in Chester and Delaware counties into South Jersey. Timing still 5:30 PM ± 1.5 hours.

Posted Thursday 06/19/25 @ 10:07 AM — I wanted to wait for a few of this morning’s models to become available before posting today.

A cold front approaches with showers and thunderstorms likely. For Philadelphia, the peak time time is 5:30 PM ± 1.5 hours. Several models are suggesting that the heaviest activity will occur just south of Philadelphia and into South Jersey.

06z HREF forecast for “maximum upward helicity”, a severity parameter, at 6 PM.

There are several signs that suggest a pattern reminiscent of last summer where the heaviest rain misses the city and its immediate suburbs.

HREF forecast of “echo tops”, the estimated height (strength) of thunderstorm clouds at 8 PM. The black areas show heights of 44,000 feet, suggesting heavy-severe storms. (Click on image for a larger view.)

We’ll have to see if that occurs.

Wed 5:54 PM —Forecast Review The severe storms did not develop as expected. Why?

The strong wave in the upper atmosphere was balanced out by strong downward motion in the lower third of the atmosphere.

Upward motion at the jet stream level (250mb)( was “neutralized” by downward motion at 5000 feet (850mb) (Click on image for a larger view.)
Strong wave at 10,000 feet (700 mb) (wavy blue contour) . (Click on image for a larger view.)
18z HRRR shows strong upward motion (violet) at 250 mb while strong downward motion at 850 mb (yellow) counteracted the strong wave. (Click on image for a larger view.)

Wednesday Storms

Posted Wednesday 06/18/25 @ 8:56 AM — Trying to determine where and when storms will develop today is like trying to decide where the first bubble will form at the bottom of a saucepan bring brought to a boil.

We have some general cues. The storms will be widespread, but a few areas may not see anything. There’s plenty of “juice” (energy) available with some models showing CAPE values as high as 3000 joules/kg. (very high). The “precipitable water” values are tropical with amounts as high as 2.7″ (extreme).

Most likely time is anytime from 1:30 PM to 6 PM with highest probability in the 4 PM to 6 PM time slot.

10z Hourly RRFS shows the heaviest rain in NJ. However, exact location is never a strong suit for weather models. (Click on image for a larger view.)

The latest hourly HRRR shows storms in the over the immediate Philadelphia area, while the latest hourly experimental RRFS shows the heaviest storms just over the river in NJ about 4-7 PM. Any storms will have extremely heavy rainfall. There will be some sunshine by late morning and early afternoon in many (but not all) areas. Any sunshine will fuel the already unstable atmosphere.


Wednesday into Thursday

Posted Tuesday 06/17/25 @ 7:41 PM — A warm front with showers and some thunderstorms is expected to move through our area later this evening into early Wednesday morning. The current radar shows the showers more eastward at this time than any of the model forecast—

Current conditions Water Vapor/Composite Radar with superimposed RAP model vertical velocity pressure (Omega- green contours) and 700 mb wind streamlines. Area of rain is about 20-30 miles more eastward than forecast by the models for this time. (Click on image for a larger view.)

Current models have the showers and thunderstorms ending early morning Wednesday, but with a range between 8 AM and 11 AM. Clouds remain for the most part, but some sunny breaks are possible in during the afternoon. Very warm and very humid conditions by the late morning.

Most models have instability showers/thunderstorms develop Wednesday afternoon between 3 PM and 6 PMs as an upper air disturbance moves through. Any sunshine will fuel their formation.

A cold front moves through Thursday late afternoon with more thunderstorms.


Tuesday into Wednesday

Posted Tuesday 06/17/25 @ 8:41 AM — Another cloudy day today. Some showers may break out this afternoon in Berks county but little in the way of showers for most of the daytime hours closer to the city.

A warm front approaches and moves through our area tonight into Wednesday morning. Around or before daybreak Wednesday, we may get some heavy thunderstorm activity that wakes everyone up a bit earlier than they’d like.

GDPS-AI forecast for Wednesday 6 AM. Approaching warm front with showers and thunderstorms. (Click on image for a larger view.)

A very humid day expected on Wednesday with dew points in the low to mid 70s!


Heat and Humidity by Wednesday

Posted Monday 06/16/25 @ 8:28 PM — The latest models reinforce the expected dramatic change in temperature and humidity in store for us. The latest model blend has temperatures approaching 90° 85° with dew points in the uncomfortable low to mid 70s as early as Wednesday. There’s a fairly high probability of heavy thunderstorms late afternoon Wednesday.

Temperatures 98°-100° still expected by next Tuesday.


Forecast Review

Posted Monday 06/16/25 @ 5:05 PM — The showers forecast to stay well west of Philadelphia this afternoon developed and moved closer to the city than forecast by most models.

Here’s the radar at 3 PM—

NEXRAD composite radar at 3 PM. (Click on image for a larger view.)

Here’s the GFS forecast rain placement—

12z GFS forecast rain placement. The light rain was forecast to be in the western suburbs. (Click on image for a larger view.)

Earlier today, I spoke of the new Canadian experimental AI modified model just released. Here’s its forecast for 3 PM today—

The new experimental Canadian GDPS- AI model did surprisingly well with today’s forecast (Click on image for a larger view.)

At the other end of the spectrum are the experimental RRFS and REFS models. I’ve held great hope in their release as operational. Unfortunately, they forecast today’s weather very poorly, with virtually no rain forecast in this time period near the city—

12z RRFS forecast for 3 PM today The RRFS forecast was unacceptable. ( Click on image for a larger view.)

They will have to become much better in order to be released. The expectation with the RRFS and REFS is for them to replace and eventually eliminate the NAM, NAM-NEST and HRRR models. I don’t see that being possible with their recent forecast errors. It would be a step backwards.


Transition to Summer Heat

Originally Posted Mon 11:15 AM — The stagnant cold, damp pattern we’ve been in over this past weekend will finally transition to summer and some significant summer heat is expected by early next week.

Current water vapor image—

Monday morning- Current satellite water vapor image with superimposed RAP model 500-1000 mb thickness lines (yellow), Mean Sea Level Pressure isobars (black contours) potential vorticity (fine violet contours) with superimposed MRMS RADAR. (Click on image for a larger view.)

We have a few more days of unsettled weather. Today (Monday) will be cloudy and warmer than recent days. A few showers are possible this afternoon, mostly in far western suburbs.

Tuesday will also be cloudy. A warm front approaches late Tuesday evening with showers and thunderstorms after midnight into early Wednesday morning.

Warmer and more humid on Wednesday with highs in the mid 80s.

A cold front moves through Thursday, also with some showers.

ECMWF-AI forecast for early Thursday morning with an approaching cold front. (Click on image for a larger view.)

Despite the “cold” front, temperatures will quickly warm over the weekend and into next week. Current AI models are predicting 95º-98º by next Tuesday.

The Canadian Meteorological Center (CMC) has recently released an AI -modified experimental (“spectrally nudged”) Global Forecasts Model at 1 hour increments and a resolution of 15 km. The model forecasts out to 240 hours at 1 hour increments up to 78 hours and 3 hour increments from 78 hours to 240 hours. Some upper level parameters are every 6 hours. This experimental model is expected to replace their regular global model, the GDPS (Global Deterministic Prediction System also called the Global GEM), in 2026.

I’ve started working on programming the download and post-processing of this model and I hope to share it here in the coming days.


Weather… and Other Things 'Up in the Air'