All Activity

Subject: 90 Years of Harrisburg (MDT) Hourly Weather — Full Deep Dive (1935–2025) Alright everyone, stuck at home in bed I went down the rabbit hole so you don’t have to. I pulled the entire hourly MDT dataset from 1935 through March 30, 2025 — roughly 701,600 hourly observations spanning just over 90 years. This wasn’t a daily summary skim. This was raw hourly temperature, dewpoint, precipitation, visibility, and related fields. Before I even looked at stats, I did cleanup and QC because raw long-term station data is messy: --- Data Cleaning & Methodology 1. File scope Period: Feb 1935 – Mar 2025 ~701,594 hourly bins Observations normalized to hourly resolution (multiple obs per hour consolidated) 2. QC steps Removed physically impossible dewpoints (e.g., below −80°F artifacts) Filtered obvious temperature spike errors (one 60°F+ 6-hour swing in 2013 was clearly a glitch) Precipitation analysis restricted to hours where precip reporting was valid (so “missing” ≠ “dry”) For daily metrics (high/low, freeze dates, heat streaks), only days with sufficient hourly coverage were used Snow depth largely unavailable in this dataset (only ~200 total depth entries), so snow retention metrics were excluded This gives us a dataset that is physically realistic and trend-consistent. --- The Results Absolute Extremes Coldest hour: −22°F (Jan 21, 1994 at 12:00) Hottest hour: 105°F (July 23, 1991 at 20:00) That’s a 127°F total observed range at MDT. --- The “Weather Whiplash” Award Largest temperature swings (QC’d): Largest 6-hour warm-up: +39.96°F (March 31, 1943) Largest 6-hour cool-down: −43.02°F (March 1, 1937) Classic Mid-Atlantic frontal violence. --- Longest Deep Freeze Continuous ≤32°F streak: 398 consecutive hours Jan 19 – Feb 4, 1961 ~16.6 straight days below freezing Runner-up: Dec 25, 2017 – Jan 9, 2018 (363 hours) So yes — modern winters can still lock in. --- Freeze–Thaw Chaos Index I counted crossings of 32°F in winter (Nov–Mar). Most chaotic winter: 1941 (188 crossings) Modern median winter: ~60 crossings That’s a real structural shift in winter volatility. --- Heat Wave Persistence (Daily-Based) Consecutive days with highs ≥90°F: Longest streak: 16 days (July 26 – Aug 10, 1983) Next longest: 15 days (July 1955) Hourly streaks don’t tell the real story because nights cool. Days do. --- The “Swamp Index” Defined as: Temp ≥90°F AND Dewpoint ≥70°F Total oppressive hours in record: 2,348 Top oppressive years: 1952 (92 hours) 1959 (87) 1949 (84) 1991 (83) Mid-century humidity was not subtle. --- Fog / Visibility Shift Hours with visibility ≤1 mile: Total: 22,368 hours ≤0.25 mile (“pea soup”): 4,860 hours By decade: 1940s–50s: ~5–6% of hours ≤1 mile 2000s–2020s: ~1.5–1.7% That is a massive drop. Possible contributors: Instrument changes Air pollution regulation Land use changes Reporting practices But the decline is statistically obvious. --- “Perfect Weather” Hours Defined as: 65–75°F Dewpoint 45–55°F No precip Visibility ≥10 miles Total perfect hours: 4,484 Best months by frequency: 1. September 2. June 3. May July is loud. September is elite. --- Precipitation Extremes Largest 1-hour precip: 4.27" (July 23, 2017) Wettest 24-hour day (reliable coverage): 16.29" (Oct 11, 2013) Wettest 72-hour total: 22.66" ending Sept 8, 2011 Those 72-hour numbers are hydrologically serious. --- Dry Spells Longest continuous dry streak (with reporting active): 286 hours Oct 20 – Nov 1, 2024 ~11.9 days Longest continuous rain streak: 33 hours (Oct 16–17, 2009) --- “Most Normal” vs “Most Weird” Year I computed daily mean deviations from long-term day-of-year climatology. Lowest RMSE (most “boring” year): 2019 Highest RMSE (most structurally weird year): 1967 1967 wasn’t just warm or cold — it was off-pattern. --- Structural Climate Signals (The Interesting Part) Comparing 1936–1959 vs 2000–2024: Diurnal Temperature Range Early period: 17.94°F average daily range Modern period: 16.36°F Change: −1.58°F Days and nights are less swingy now. Night vs Day Warming Mean daily minimum change: +2.64°F Mean daily maximum change: +1.07°F Nights warming more than days is a classic humidification / greenhouse signature. Dewpoint Drift +0.79°F increase in daily mean dewpoint Combine: Warmer nights Slightly higher dewpoints Reduced diurnal range That suggests a shift toward a more moisture-retentive boundary layer and fewer extreme radiational cooling setups. --- Growing Season Length Using freeze date methodology: Earliest first fall freeze: Oct 16, 1939 Latest first fall freeze: Nov 13, 2024 Longest growing season: 233 days (2024) Shortest: 158 days (1983) Typical: ~203 days The envelope is expanding. --- Big Takeaways 1. The 1940s–1960s were much foggier and had more freeze–thaw volatility. 2. The 1980s still dominate heat wave duration. 3. Extreme precipitation events are clustered in the 2000s–2010s. 4. Nights have warmed faster than days. 5. Diurnal range compression is measurable. 6. Dewpoints are nudging upward. 7. Growing season envelope has widened. This isn’t just “it’s warmer now.” It’s structural atmospheric behavior shifting. Sent from my SM-S731U using Tapatalk

Gfs

Damn, you make @mitchnick look young.

“Our”? Seems ok to me.

Today was a pack destroyer down here, no doubt. Lucky that it was such a moisture-packed glacier, holding onto about 3-5 inches of actual pack. Piles still going strong, that’s gonna take a lot more to melt through. 9-10 days have been above freezing this month so far, just really excellent retention this year - the deep freeze damn near turned it all to stone. Back down to 28f.

Growing up in central NJ, 60-61 was unbelievable. Three major snowstorms of a foot plus, The first, starting on a Saturday night (maybe Dec 9th) and ending Sunday closed schools for the whole week. The other two storms were in Jan and early Feb. Afterwards, I wondered why we did not get storms like that every year.

My wife and I headed up to Bolton Valley for a lift-served session this morning, so I can pass along the conditions updates. We planned to spend our time on Timberline and arrived there right around the opening of the Timberline Quad. Being a holiday weekend, we were a bit worried about how busy the resort was going to get, but the Timberline Quad remained at pretty much walk-on levels of traffic right up through noontime when we were done. So, skier traffic was relatively light, but the snowfall from the larger storm earlier in the week was really spread out over the course of multiple days (Tuesday through Thursday). So, with respect to lift-served terrain, that storm’s powder has certainly been skied. Therefore, on piste areas and easily accessed off piste areas are mostly packed at this point, and we had to travel a bit farther afield to get into untracked lines. We did pick up another inch or two of snow today down at our place in the valley, and it snowed lightly all morning on the mountain to add similar amounts, so that added another layer of freshened up snow. In general, the powder was in excellent shape, albeit with a bit more settling that what I encountered on my ski tour yesterday. We actually encountered a bit of sun/melt crust on some southerly aspects down around 1,500’, so be aware of that possibility if you’re planning to hit a any low elevation terrain with substantial southerly exposure. On piste conditions were excellent packed powder aside from the usual areas that have a manmade base, so it looks like the holiday weekend will continue to roll on with great midwinter conditions in the Northern Greens.

From 54 to 34 and dew does not rise

This storm is crap

yep I lost 3 plus inches in my front yard today

-Slightly on the lower side of climo snowfall for PWM, but season-to-date is still considered AVG. -Relatively persistent cold ---2 weeks in Jan AN but not a torch and no f*kn cutters. -Plenty of days with sun and little or no wind. Oh boo hoo it was only 12°. shh. -P-sure Xmas was white, too.

Everything going to shit Not our winter

Icon has no snow

I want 80 more at SLK lol

Temps 22-54 today, with sun. Significant range. Did some shoveling , but ice on the north side remains impenetrable even though meltwater ponded on top. Have been using old rubber mats and wood ash to keep walking paths safe. Not much longer I guess .

She sold her soul to the junk man….

It just snowed south of Tampa. Islip NY just had 19 consecutive days with a low temp of 19° or colder, the longest stretch on record, going back to the 1800s. It snowed 10" in Florida last Winter, which was the most on record for the state since the 1800s, previous record 4". There is meteorological reasoning, it's not because Donald decided to drive his car today. I'm not arguing with you global warming in general: There are macro states that explain it well. I think it's more about you complaining and not wanting to concede that this was predicted 8 days ago, by methods you disagreed with. I'm not really engaging in your tangent about "where have all the borderline storms gone". I said it before, and I'll say it in the future, the Pacific with +450dm is extreme, and a fast Pacific jet disconnects the northern stream, and floods the US with low level warm air. A storm like this gets cutoff from the freezing line. It's not a borderline event at all: Our average high temp in the heart of Winter in strong -PNA is mid 40s. I don't care if 6% of the examples broke the trend. I know in 97-98 we had these "perfect track rainstorms" all the time.. I think 72-73 too. etc. I know the average statistics, and you are riding hard on anomalies, it usually won't work out.

Here are some that I got. One is the lunar eclipse from March 2025...if you look closely, those are actually stars to the right in the photo. The normally bright full moon would have totally obscured those, but during a total lunar eclipse, it is very muted. The others include: a double rainbow, an incoming storm near sunset, blood-red clouds near sunset, and mammatus clouds at the edge of a storm (maybe from the same storm that @Bob Chill showed above??). I always like the sunset colors, the sky and any clouds change color constantly.

I'm starting to catch your angst over this and I don't like it, lol I'm gonna just pretend a better PDO will save the day...and or simply believe the next Niño will finally deliver our HECS

I'm kinda impressed at how well the Euro AI did: https://www.tropicaltidbits.com/analysis/models/?model=ec-aifs&region=us&pkg=mslp_pcpn&runtime=2026020718&fh=192

maybe lock it until the first flakes lol

Today's split EWR: 47 / 28 (+3) NYC: 46 / 29 (+2)

My suggestion...ride that thing into the ground until jeep prices come down. Those engines and transmissions can last unless you're off-roading on the regular.

Lock the thread, and models come north, unlock thread models go south ... Hmm