Sign in

Jonathan Lee

@lee-cyclone.bsky.social
36 followers 101 following 34 posts

Atmospheric science PhD student at NYU | Cornell ‘25

PostsRepliesMedia
Jonathan Lee @lee-cyclone.bsky.social · 25/08/2026
one could use emergent constraints but that might not be the intent here
010 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
Some thoughts (here I implicitly assume a fixed threshold definition) - bsky.app/profile/lee-...
011 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
(None of this is particularly new, but perhaps a useful and underappreciated insight - and a relevant one considering the increasing frequency of extreme weather in recent days and decades.)
000 0
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
This means that some compound extremes could be increasing at faster rates than the sum of individual hazards alone, explained by (a) simple probability (assuming each hazard’s own probability is rising), (b) changing physical relationships between two hazards (especially if strengthening).
100 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
… or there is a common driver behind two different hazards (e.g., a persistent jet stream/block enabling both heat and flooding in different places).
100 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
Oftentimes different physical hazards are linked in some way (via feedbacks or couplings, or one hazard often serves as a precursor to another (e.g., in some areas, extreme heat —> more instability and moisture —> added flooding risk) …
100 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
The independence assumption above is naive. Even in a past climate, the compound event probability could be greater (or less than) ab.
100 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
Example: If A and B each become 20% more frequent, compound A-B events become 44% more frequent.
100 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
Now suppose warming shifts each individual risk, so that the new probabilities are a+h and b+k. The (independent) joint probability of both occurring at the same time is (a+h)(b+k) = ab+ak+hb+hk. (This is more than linear/additive — relative changes multiply rather than add.)
100 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
Consider two hazard types A and B (e.g., heat and flooding) occurring with probability a and b, respectively — in a past climate, for a given region. Assuming independence for now between the events (which may or may not hold - more later), the combined probability of occurrence is ab.
100 1
Jonathan Lee @lee-cyclone.bsky.social · 19/07/2026
🧵Compound extremes are probably increasing faster than individual extremes in a warming world: a thread
140 4
Jonathan Lee @lee-cyclone.bsky.social · 07/07/2026
(5/5) Also it is worth noting that, broadly, climate change is intensifying extreme precipitation events, and most so in the Northeastern US. (Ref. www.nrcc.cornell.edu/services/blo... and repository.library.noaa.gov/view/noaa/32...)
Observed Changes in the Frequency and Severity of Heavy Precipitation Events. In the Northeast, total precipitation on the heaviest 1% of days has increased ~60% in the past few decades.
020 2
Jonathan Lee @lee-cyclone.bsky.social · 07/07/2026
(4/5) Impressively, a nearly stationary cell persisted just south of Queens for hours. Here is a radar loop of the rainfall showing that. There is a secondary area of 2”+ rain well north of NYC, also due to a persistent, quasi-stationary band of showers.
100 1
Jonathan Lee @lee-cyclone.bsky.social · 07/07/2026
(3/5) A “moderate” risk (threat level 3 out of 4) of flash flooding was advertised for the NYC and LI area. The potential for flash flooding was there, but the exact location of heavy rain was (and often is) hard to pinpoint much in advance, though.
100 1
Jonathan Lee @lee-cyclone.bsky.social · 07/07/2026
(2/5) Fortunately the bullseye of the rain was offshore, though southern Queens and Brooklyn got quite a bit of rain. Many areas got 2”+ of rain, though Manhattan <1”, with isolated spots up to 4” in Long Island.
100 1
Jonathan Lee @lee-cyclone.bsky.social · 07/07/2026
🧵(1/5) Just ~15 miles southeast of midtown Manhattan, it rained about 8” in the past 24 hours. It would have been a vastly different story had storms were just a few miles north …
Map of past 24 h rainfall in NYC and Long Island area. Many areas got 2” of rain with a spot just offshore getting up to 8”.
100 1
Jonathan Lee @lee-cyclone.bsky.social · 03/07/2026
yeah, I would guess that the actual value of c < 0.5
010 1
Jonathan Lee @lee-cyclone.bsky.social · 03/07/2026
Using w ~ c*sqrt(2*CAPE) and CAPE ~ 3000 J/kg (from SPC mesoanalysis), and assuming c ~ 0.5 (a very rough estimate), w ~ 39 m/s, which seems quite high for a pulse storm
110 1
Reposted by Jonathan Lee
Andrew Dessler @andrewdessler.com · 30/06/2026
On The Climate Brink, I explain how we connect extreme weather to climate change. Yes, it's another draft chapter from my forthcoming climate risk textbook. www.theclimatebrink.com/p/how-climat...
theclimatebrink.com
How climate change influences extreme weather
How science shows us how warming makes heat waves, floods, hurricanes, and wildfires more likely or more severe — and why this science is becoming a major public-policy battleground.
23214 32
Reposted by Jonathan Lee
Katharine Hayhoe @katharinehayhoe.com · 28/05/2026
RCP and SSP scenarios are not emission scenarios. I repeat: they are not emission scenarios. They are radiative forcing scenarios. Forcing = human emissions + the response of the climate system to those emissions You can only back out emissions by making assumptions about climate sensitivity.
220656 206
Jonathan Lee @lee-cyclone.bsky.social · 03/04/2026
and today’s temperatures - a remarkable warmup especially in upstate NY compared to yesterday, with a cooler coast (almost opposite Wednesday’s temperature pattern).
upper 70s in northwest NY, otherwise widespread 60s to low 70s. NYC and LI 50s to mid 60s.
010 1
Jonathan Lee @lee-cyclone.bsky.social · 02/04/2026
Early April temperatures in New York State: yesterday vs today. Some remarkable temperature gradients!
Apr 1: temperatures in mid 30s to mid 40s in northern half of the state, and mostly 60s and 70s downstate (lower Hudson Valley and NYC)
 Temperatures in the mid 30s to mid 40s now covering the eastern half of NYS (including NYC and Hudson Valley, which were in the 70s). But extreme western part of NY is in the 60s and low 70s, with a sharp gradient.
110 1
Reposted by Jonathan Lee
Daniel Swain @weatherwest.bsky.social · 24/03/2026
Given recent discussion regarding indirect role of anomalous precip (& associated upstream diabatic heating) in amplifying record-breaking heatwaves, even thousands of miles away (inc. Mar 2026 event), I wanted to share this excellent paper from 2025:
nature.com
Blocking diversity causes distinct roles of diabatic heating in the Northern Hemisphere
Nature Communications - Atmospheric blocking patterns are often categorized into several types and diabatic heating is thought to primarily contribute positively to blocks. Here, the authors show...
28718 87
Reposted by Jonathan Lee
Pam Kan-Rice @ucanrpam.bsky.social · 19/03/2026
Phone weather apps are inaccurate. Those apps are fully automated and use algorithms that aren’t “sufficiently dynamic,” @weatherwest.bsky.social @ucanr.edu said — and in a nutshell, they’re lacking human expertise and customization behind the scenes.
45825 58
Reposted by Jonathan Lee
Tomer Burg @burgwx.bsky.social · 12/03/2026
One of the wildest 24-hour loops you'll see in the Mid Atlantic - from mid-upper 80s and tornado warnings to 33F and heavy snow
04920 49
Jonathan Lee @lee-cyclone.bsky.social · 26/02/2026
AI/ML algorithms can sometimes generate unphysical forecasts: a case in point
A forecast from Apple Weather showing snow next Thursday to Saturday, despite mild temperatures (37/46, 42/51, 43/61 F).
000 0
Jonathan Lee @lee-cyclone.bsky.social · 02/02/2026
Here is the (more reliable) seasonal temperature forecast from NOAA’s Climate Prediction Center.
Above normal temperatures likely for the southern tier of the US, below normal likely in the Northern Plains and Northern Rockies, equal chances elsewhere.
000 0
Jonathan Lee @lee-cyclone.bsky.social · 02/02/2026
Over the past decade, Phil has only been accurate 30% of the time in subseasonal forecasts of spring temperature anomalies in the US. More details in the attached abstract (and graphical abstract, courtesy of NOAA).
Phil saw his shadow 108 times (“more winter”), no shadow 21 times (“early spring”, and no record was made 10 times (since 1887). But he has only been right 30% of the time in the past 10 years. 17.1 - Evaluation of a Simple Terrestrial-Based Predictor for Subseasonal Prediction of Early-Spring Onset 
Phil, P. 

February 2, 2026, 7:30am-7:45am
Gobbler’s Knob (Punxsutawney, PA)

Research in and demand for subseasonal-to-seasonal forecasts have been surging in recent years. Of particular interest is the phenomenon of early-spring onset and whether it can be predicted on subseasonal scales. We present a simple terrestrial, marmot-based prognostication model which only depends on the presence of the shadow of a particular M. monax as it emerges from its burrow, where the absence of a shadow is hypothesized to bring about an early-spring onset. Forecasts are initialized on February 2 of each year between 7:00 and 8:00am Eastern Time (around 12-13 UTC), with an observational record dating back to 1887. Preliminary analyses reveal that the model outputs a shadow and thus favors no early-spring onset in most years, and that forecasts in the past decade have yielded roughly a 30% accuracy. It is concluded that a groundhog’s shadow does not provide meaningful skill to predictions of early-spring onset on subseasonal time scales relative to climatology.  

140th Special Session on Groundhog Meteorology
100 1
Reposted by Jonathan Lee
Simon Lee @simonleewx.com · 23/01/2026
The extreme cold across eastern N America is due to the Alaskan Ridge regime, which on average brings the coldest winter weather to this region. Is this happening more often? No. During the modern warming era, there has been no observed trend in the occurrence of this regime at this time of year.
A time series of the January–February frequency (in days) of the AKR regime from 1979 to 2025 using ERA5 data, based on the year-round classification of Lee et al. 2023 (J. Climate) updated through 2025.
38028 80
Jonathan Lee @lee-cyclone.bsky.social · 11/01/2026
Given that SEAS5 seems to capture well the extreme frequency/duration of GH blocks, if the GH persistence trend happens to be caused by internal variability (or eventually stops), should one still worry about a systematic bias in the seasonal and/or climate models?
100 1
Reposted by Jonathan Lee
Jonathan Lee @lee-cyclone.bsky.social · 24/12/2025
🧵How might snowfall change in a warming world? I set up a very simple “toy” statistical model to explore this question. Isn't it obvious that snowfall decreases as temperatures get warmer? Yes, but …
142 4
Jonathan Lee @lee-cyclone.bsky.social · 24/12/2025
🧵How might snowfall change in a warming world? I set up a very simple “toy” statistical model to explore this question. Isn't it obvious that snowfall decreases as temperatures get warmer? Yes, but …
142 4
Jonathan Lee @lee-cyclone.bsky.social · 24/12/2025
More details and plots found here: colab.research.google.com/drive/18sc0I.... The plots have interactive sliders, and you can change the temp and other parameters. (Works best on laptop or tablet).
colab.research.google.com
Google Colab
000 0
Jonathan Lee @lee-cyclone.bsky.social · 24/12/2025
How does total snowfall depend on the mean and stdev of temp? As spread of T increases, the “optimal” temp for snowfall decreases (since more snow is lost to the warmer tail), from ~-2° to -8°C. At cold enough T, warming increases snowfall modestly (⬆️ moisture dominates).
Total snowfall (arbitrary units) as a function of standard deviation and mean temperatureTotal snowfall sensitivity (% change per C warming) as a function of standard deviation and mean temperature.
110 1
Jonathan Lee @lee-cyclone.bsky.social · 24/12/2025
With some simplifying assumptions, I use two equations to estimate the change in snowfall (more details in link later) Here I show how warming from -6±5°C by 1 degree changes snowfall (intensity-weighted frequency). κ = 1.07 (7% more intense per C), c = 1 (Clausius-Clapeyron scale factor)
Graph of intensity weighted snowfall frequency vs temperature. Total change in snowfall is -0.92%, with loss from less cold -6.47% and gain from more moisture +5.54% (amounts may not add up exactly due to rounding error) for kappa = 1.07, or 7% moisture increase per degreeGraph of intensity weighted snowfall frequency vs temperature. Total change in snowfall is -0.92%, with loss from less cold -6.47% and gain from more moisture +5.54% (amounts may not add up exactly due to rounding error)  for c = 1.0 (Clausius-Clapeyron scaling)
100 1
Jonathan Lee @lee-cyclone.bsky.social · 24/12/2025
… precipitation intensity does depend on the temperature. A warmer atmosphere can “hold” more moisture (roughly 7% per °C). ⬇️ frequency may be compensated by ⬆️ moisture. How important is each effect? When does warming actually *increase* snowfall?
100 1
Jonathan Lee @lee-cyclone.bsky.social · 18/12/2025
The success of atmospheric science in recent decades was built on steady, gradual improvements - a silent revolution - as computer models (like WRF, developed at NCAR) advanced & forecast errors declined. NCAR has played an instrumental role in the development of the entire field. #SaveNCAR (3/3)
“Predicting Weather” exhibit at NCAR. Models are used to predict weather in the near future.
Drawing on information about current weather, and applying what we know about how the atmosphere works, weather models predict what will most likely happen. For example, while we know the temperature now, a model is programmed to consider and account for all the things that will affect temperature in a few hours to a few days.
Models are used to explore and experiment.
What would happen if a hurricane moved over warmer ocean water? What would happen if a cold front moved quickly into a warm front? Looking at questions like these, scientists can use a weather model to test how the atmosphere would react under hypothetical conditions. The same weather events happen in the real world too, but in a model scientists can experiment with the weather. They can also run the model many times to see what factors affect the forecast and to determine its accuracy. / Weather forecasts have improved over the last 50 yecrs, largely as a result of improving weather models.
How accurate can forecasts become? NGAR scientists improve forecasts by developing new models that better describe the real atmosphere. Increasing computing power mokes models more detailed and more accurate. Finally, many researchers are creating new ways to bring data from satellitos, radars other observing systems into models. This helps give the mode's a better "head start" to predict future weather. NCAR scientists work with scientists all over the world to develop and improve the Weather Research and Forecasting model (WRF). This is a freely available, widely used forecasting tool, Millions of people worldwide rely on weather forecasts developed using the WRF model.
051 5
Jonathan Lee @lee-cyclone.bsky.social · 18/12/2025
Here’s a video of an interactive tornado exhibit there! (2/3)
120 2
Jonathan Lee @lee-cyclone.bsky.social · 18/12/2025
#NCAR is a beautiful place and a national treasure. I had the chance to visit the NCAR Mesa Lab last summer, and their exhibits (free and public) are the coolest and most informative ones I’ve seen on atmospheric and climate science. (1/3)
The front entrance of the sandstone colored NCAR building, with the words “National Center for Atmospheric Research / Sponsored by the National Science Foundation”.
An exhibit titled “Change Over Time”. Climate has changed over time, not only in regions, but also for Earth as a whole. Global temperature has risen and fallen many times, often in response to changes in how the Earth tilts on its axis and orbits the Sun. Large volcanoes and meteors produce temporary global and regional cooling.
Earth's temperature has risen more than 1.3°F (0.7°C) in the last century and most dramatically since the 1970s. The burning of fossil fuels has increased the amount of greenhouse gases in Earth's atmosphere each year since measurements began in the 1950s.
Carbon dioxide and other greenhouse gases absorb heat energy from Earth, allowing less to escape to space. Increasing the amount of these gases makes Earth warmer. Our planet is expected to warm even more if fossil fuel use continues.
Climate models project that many parts of the world, including North America, will experience changes this century. These include more frequent and severe heat waves and more intense bursts of rain and snow with longer dry periods in between. Some of these trends are already being observed.
Humans have always been able to adapt to weather and to changes in climate. Now we are beginning to face the challenge of adapting to a rapidly changing climate with conditions very different than in the past.“A history of understanding” exhibit, with portraits of five individuals (Fourier, Foote, Tyndall, Arrhenius, Callendar). For thousands of years, Indigenous cultures around the globe have viewed the land, sea, and sky as interconnected. Scientists have also known for about two hundred years that carbon dioxide and other greenhouse gases in the sky are formed with elements from the land and sea and affect Earth's temperature. In the 1820s, French scientist Joseph Fourier proposed that the atmosphere might be insulating our planet - the first inkling of what we call the greenhouse effect. In 1856, American scientist Eunice Newton Foote discovered that carbon dioxide and water vapor absorb heat in sunlight. She suggested that changes in the amounts of these gases could change the climate. Three years later, British scientist John Tyndall tested many gases and found that some were able to absorb and emit heat, confirming the greenhouse effect.
In 1896, Svante Arrhenius calculated how increasing carbon dioxide from burning fossil fuels could warm our planet. He speculated that climate warming might be a good thing if it could help farms produce more food, stave off the next Ice Age, and improve the climate of places like his native Sweden. In the 1930s, British steam engineer G.S. Callendar demonstrated that burning fossil fuels was warming the planet. We now know that rapid warming is leading to many problems and uncertainties across the world.An artwork (painting?) of the NCAR Mesa Lab building with mountains and (presumably) cumulonimbus clouds in the background.
183 8
Jonathan Lee @lee-cyclone.bsky.social · 13/12/2025
agreed, and this might inspire me to share some mediocre GFD jokes (yes, I presented a GFD-themed stand-up comedy at a student seminar)
010 1
Jonathan Lee @lee-cyclone.bsky.social · 27/02/2025
Sorry to hear this, Zack.
010 1