Deciphering Changing Probabilities of Extreme Climate Events in Climate Models and Measurements

Background

Due to anthropogenic activities since the Industrial Revolution, the earth’s climate has been experiencing substantial changes. From 2011 to 2020, global surface temperature has reached 1.1 °C above pre-industrial levels (GISTEMP v.4 2024) and is projected to rise. The increase in global temperature will throw off the equilibrium of the earth's systems, creating environmental, societal, and economic impacts.

Human-induced climate change is increasing the frequency and intensity of extreme climate events in every region of the world, including those related to precipitation. Under a changing climate, weather will be warmer and thus wetter. This phenomenon is supported by the Clausius-Clapeyron equation, which describes how the saturation vapor pressure of water increases exponentially with temperature, leading to more moisture in the atmosphere and potentially heavier precipitation events (Martinkova et al. 2020). Future climate is likely to have an increase in both frequency and intensity of precipitation events (Myhre et al. 2019).

Since the 1950s, most regions of the world have experienced increases in hot extremes, 21 out of 46 global regions experienced an increase in heavy precipitation, and West and Central Africa as well as parts of Europe and Asia experienced more occurrences of drought (IPCC AR6).

Understanding the shifting probabilities of extreme climate events is incredibly important to prepare the world for the escalating occurrences of severe weather phenomena. As precipitation patterns shift, it is paramount for us to be ready for both extreme precipitation events (which can have consequences such as severe flooding, landslides, infrastructure damage, and proliferation of waterborne diseases) and droughts (which can have consequences such as water supply shortages, agricultural impacts, and increased risk of wildfires).

A High Level View of Precipitation Patterns

How Precipitation Patterns are Changing in the United States

As global temperature increases, there will be an increase in annual hottest-days and wettest-days (IPCC AR6). The frequency of extreme precipitation events is projected to increase significantly with global warming, with total precipitation from these events nearly doubling per degree of warming, primarily due to increased frequency rather than intensity. This trend, observed in both climate models and actual observations, suggests that the most intense precipitation events will become almost twice as common with each additional degree of global warming (Myhre et al. 2019).

A precipitation return value is a statistical estimate representing the magnitude of a precipitation event that is expected to occur once in a given period. For example, a 100-year return value (sometimes called the '100 year storm' or 'once in a 100 year storm') refers to an extreme rainfall event that has a 1% chance of happening in any given year. Since the 1950s, precipitation return values have generally increased in the Eastern United States and decreased in the Western United States (Wu 2015).

Figure 2: Precipitation Return Values across the contiguous United States based on Weather Station Data over Several Decades (1955-2015) (Risser et al. 2019)

Findings from the 2023 National Climate Assessment (NCA-5):

• Extreme precipitation patterns in the United States are undergoing significant changes due to climate change. The intensity and frequency of the heaviest precipitation events are increasing, posing heightened risks to both people and nature.
• Regions like Alaska and the northern and eastern parts of the U.S. are experiencing, and are expected to continue experiencing, more average precipitation, whereas areas such as Hawai'i and the southwestern U.S. are seeing declines. These shifts contribute to an increase in heavier rainfall events nationwide, with very high confidence that these trends will continue.
• Increased temperatures are expected to enhance evaporation and plant water use, especially where moisture is plentiful, further stressing groundwater supplies.
• As snow cover diminishes and melts earlier, these changes threaten freshwater availability, exacerbate flood damage, and increase drought impacts.
• These changes in precipitation patterns particularly affect socioeconomically disadvantaged communities and those on the frontlines of climate change.
• The combined effect of more frequent and intense precipitation events necessitates rapid adaptation to mitigate the associated flood and drought-related challenges.

Climate Models and Precipitation

Climate models are essential for assessing climate variability and designing infrastructure, but they may not accurately predict precipitation extremes, which are expected to increase in frequency and intensity due to climate change (Abdelmoaty et al. 2021). Climate models often struggle to accurately predict precipitation extremes due to several inherent biases and limitations. One key issue is the "drizzling bias," where models tend to simulate precipitation events that are too frequent and prolonged but with lower intensity compared to observations. This results in an overestimation of light rain events and an underestimation of heavy precipitation events. Coarser models exacerbate this problem as they rely heavily on parameterized convective precipitation, which is more drizzle-like (Chen et al. 2021).

Working with NASA's GISS ModelE2.1, a coarser resolution general circulation model (GCM) amongst the CMIP6 group (Miller et al. 2020), we found differences between the model's representation of precipitation distribution and observational data. We used IMERG v.7, a remote sensing product distributed by the NASA Global Precipitation Measurement Mission, as our observational comparison (GPM_3IMERGHHE).

Figure 3: Comparative Analysis of Precipitation Distribution: The NASA GISS Climate Model vs. IMERG v.7 Remote Sensing Data
Compared to sattelite retrievals from IMERG v.7, the GISS climate model (ModelE2.1) does not capture the most extreme of precipitation in New York City 2011-2020. While the GISS model's precipitation values had a slightly higher average and 95th percentile than IMERG for this region, the observed IMERG precipitation values had a 25% higher 99th percentile value than GISS. The IMERG data used in this plot was regridded to the coarser GISS grid for comparison at the same resolution.
The source code for the regridding, preprocessing, and plotting used in this figure is available in this project's Github repository in the 'Probability Density Functions' folder.

Extreme Precipitation Events

Having looked at extreme precipitation from a total distribution standpoint, we moved forward in the project examining individual precipition events. To start, we examined IMERG v.7 precipitation data for the New York City area from 2001 to 2022 to find precipitation events and rank them by total precipitation, precipitation rate, and precipitation events with the highest individual measurements. The source code used for this analysis is available in this project's Github repository in the 'Top Precipitation Events' folder.

After we identified 100+ top precipitation events in New York City within IMERG v.7, we looked at the events individually by identifying if the event was in a National Weather Service (NWS) or NOAA database, if the event was declared a disaster by FEMA, and if the event had substantial news coverage. During this process, we discovered that many of the events we identified were Nor'easters (including this one on December 5, 2020) and Bomb Cyclones (including this one on October 17, 2019). We then shifted our research goal to focus on Bomb Cyclones (described below) and winter precipitation.

Extratropical Cyclones (ETCs)

An extratropical cyclone is a large-scale weather system that forms outside the tropics, typically featuring a low-pressure center, cold and warm fronts, and associated weather patterns such as strong winds, rain, and sometimes snow. Extratropical cyclones are a major component of mid-latitude precipitation in the Northern Hemisphere, accounting for approximately 80% of winter precipitation in this region (Naud et al. 2022). ETCs which undergo bombogenesis, or ‘bomb cyclones’, account for particularly heavy precipitation (Seiler et al. 2016). There is a lack of consensus on how extratropical cyclones (ETCs) will change under a warming climate (Naud et al. 2022), and the majority of CMIP5 climate models struggle to accurately predict the frequency of bomb cyclone occurrence (Seiler et al. 2016). Additionally, ETCs have stronger effects when they are concurrent with Atmospheric Rivers, or ARs, (Zhang et al. 2019) and the frequency and intensity of ARs is predicted to increase under the changing climate (Zhang et al. 2024).

Figure 4: Recent News Stories about Extratropical Cyclones and Bomb Cyclones impacting the U.S. East Coast. Click the images to read more.

Bomb Cyclones

Bomb Cyclones are extratropical cyclones which undergo rapid intensification due to a rapid drop in pressure at their storm center. Bomb Cyclones are generally characterized by a pressure drop of at least 24 millibars within 24 hours, but this pressure drop threshold is latitude dependent and may be less than 24 millibars in some cases (Sanders and Gyakum 1980). Bomb Cyclones can produce severe weather conditions, including heavy snowfall, strong winds, and coastal flooding, making them particularly dangerous and impactful. They often originate from the western North Atlantic in the winter months.

Figure 5: An Example of a Bomb Cyclone (source: weather.com).
This animation shows a cold air mass (high pressure system) moving southeast while a warmer air mass (low pressure system) moves northward. The steep gradient in temperature and air pressure created a dangerous bomb cyclone on this date.

Atmospheric Rivers

An atmospheric river is a narrow corridor of concentrated moisture in the atmosphere that can transport vast amounts of water vapor from tropical regions to mid-latitudes, and they are often associated with extratropical cyclones (LeGrande et al. 2024). ARs can stretch for thousands of kilometers and are often responsible for significant rainfall when they make landfall, potentially leading to flooding. Atmospheric rivers play a crucial role in the global water cycle but can also cause severe weather impacts, such as heavy precipitation and strong winds (NASA GHRC). Atmospheric Rivers (ARs) are most often associated with western coastal regions, but they also have impacts beyond those regions, including across the U.S. East Coast (O'Brien et al. 2024). Both extratropical cyclones and atmospheric rivers tend to be stronger when they are combined (Zhang et al. 2019). Approximately 80% of cyclones in North Atlantic and North Pacific basins that undergo bombogenesis (i.e. bomb cyclones) are linked to the presence of an AR; in contrast, less than half of cyclones that do not undergo bombogenesis (i.e. non-bombing cyclones) are associated with an atmospheric river (Eiras-Barca et al. 2018). There is projected to be a 50% increase in AR occurrence globally by the late 21st century (Payne et al. 2020).

Figure 6: An Atmospheric River moves Water Vapor and Precipitation from the Tropics to the U.S. West Coast (source: JPL).
This visualization uses satellite data to show the movement of water vapor and precipitation as an atmospheric river slams into California.

Research Question

With a focus on the U.S. East Coast, we aimed to investigate how the frequency and intensity of extratropical cyclones, and specifically bomb cyclones, are represented in various datasets (including reanalysis and climate models) and how their characteristics are changing under climate change. Additionally, because ETCs have stronger effects when they are concurrent with atmospheric rivers, (Zhang et al. 2019) and the frequency and intensity of ARs is predicted to increase under the changing climate (Zhang et al. 2024), we also aimed to investigate atmospheric concurrent ETCs and BCs.

Back to the table of contents

Methods

Datasets and Tools

• NASA Climate Model: GISS-E2.1 (Kelly et al. 2020)
  • 2°x2.5°x40-layer atmosphere (to 0.1mb)
  • Nudged to MERRA-2 horizontal winds
  • 6-hourly instantaneous values
• Observational Datasets:
  • Weather station data (Risser et al. 2019, National Weather Service)
  • IMERG v.7 (remote sensing): GPM IMERG Final Precipitation L3 Half Hourly 0.1 Degree x 0.1 Degree V07. doi:10.5067/GPM/IMERG/3B-HH/07
  • NOAA daily weather maps
• Reanalysis Datasets:
  • MERRA-2: Global Modeling and Assimilation Office (GMAO)(2015), MERRA-2 tavg1_2d_flx_Nx: 2d,1-Hourly,Time-Averaged,Single-Level,Assimilation,Surface Flux Diagnostics V5.12.4, Greenbelt, MD, USA, Goddard Earth Sciences Data and Information Services Center (GES DISC), doi:10.5067/7MCPBJ41Y0K6
  • ERA-5 (hourly data on single levels): ERA-5: European Centre for Medium-Range Weather Forecasts, Copernicus Climate Change Service, ERA5 hourly data on single levels. doi.org/10.24381/cds.adbb2d47
• Pre-processed Data Products:
  • Dataset of ETCs identified in MERRA-2 reanalysis using MAP Climatology of Mid-latitude Storminess (MCMS), 2007-2021 (Booth et al. 2013, Provided by Catherine M. Naud)
  • Dataset of ETCs identified in ERA-5 reanalysis using MAP Climatology of Mid-latitude Storminess (MCMS), 1950-2023 (Booth et al. 2013, Provided by James Booth)
• Tools and Software used for Analysis:
  • Python
  • Panoply (NASA's geospatial data visualization software)

Method for Identifying Extratropical Cyclones, including Bombogenesis and Atmospheric River Concurrency, in Gridded Datasets

Identifying Extratropical Cyclones

Our first method involved capturing extratropical cyclones identified in MERRA-2 and ERA-5 reanalysis using the MAP Climatology of Mid-latitude Storminess (MCMS) method (Booth et al. 2013) which passed through our U.S. East and West Coast region masks in order to organize ETCs by relevant region. The global datasets of ETCs identified with MCMS were provided to us by Catherine M. Naud (MERRA-2) and James Booth (ERA-5) (Booth et al. 2013). U.S. East and West Coast ETCs were captured from 2007-2021 in MERRA-2 and 1950-2023 in ERA-5.

Figure 7: U.S. East and West Coast Region Masks. Coastal boundaries are based on RECCAP2 (Laruelle et al. 2017) In this plot, the white grid boxes represent land, and the purple grid boxes represent coastal ocean. These mask files are available in this project's Github repository.

Identifying Bombogenesis

After regionally relevant ETCs were identified, each of the ETCs’ storm paths were searched for pressure drops associated with bombogenesis as defined by the latitude dependent equations set forth by Sanders and Gyakum 1980. Pressure drops were considered at the storm’s center and within 4 degrees orthogonal and diagonal to the storm’s center, all along the storm’s path. If the pressure drop threshold was met during a 24 hour time period at any center point along the storm's track, the ETC was marked as “bomb”; if no pressure drops that met the threshold were found, the storm was marked as “non-bomb”.

Identifying Atmospheric River Concurrency

The area around each ETC was checked for Atmospheric Rivers to determine AR-concurrency. Identified atmospheric rivers were sourced from GISS ModelE2.1 nudged to MERRA-2 horizontal winds (LeGrande et al. 2024), but in the future should be sourced from ERA5 or MERRA-2 (Guan and Waliser 2024, database of ARs available on UCLA Dataverse). A composite of an AR-concurrent ETC from Zhang et al. 2019 was used to determine a proximity box in which ARs were associated with the ETC in question. If an atmospheric river existed within the proximity box identified at any timestep of the storm, the storm was marked as 'AR-concurrent'.

Figure 8: Atmospheric River Concurrency - Proximity Box (Zhang et al. 2019). The probability (colors, %) of atmospheric river (AR) occurrence around the composite extratropical cyclone (EC) center (black dot). The red square shows the proximity box used to identify Atmospheric River concurrency in our project.

Identifying Storm-Specific Precipitation

We developed a method to identify storm-specific precipitation by finding the Sea Level Pressure (SLP) closed loop farthest away from the storm center which contains the storm center. To achieve this, we first loaded the SLP data and used the Haversine formula to calculate the distances between the storm center and each grid point in the dataset. We then identified the grid point closest to the storm center.

Next, we identified closed SLP loops by labeling contiguous areas of relatively low SLP values. Specifically, we considered regions where SLP values were below the median (50th percentile) of all SLP values in the dataset. This threshold allowed us to identify a manageable number of closed loops while ensuring that these loops represented significant low-pressure systems. We then calculated the centroid of each closed loop to determine their central coordinates. For each loop, we checked if the storm center was within the loop's boundaries.

Among the loops containing the storm center, we selected the one with the centroid farthest from the storm center by calculating the distances from the storm center to the centroids of these loops. We then calculated the area of this selected loop and created a mask corresponding to this loop, which we used for subsequent precipitation analysis. This mask allowed us to isolate and study the precipitation associated with the identified closed SLP loop.

Figure 9: Identifying Storm-Specific Precipitation Example of December 26, 2010 AR-Concurrent Bomb Cyclone
The figures show Sea Level Pressure (SLP) and precipitation data from MERRA-2 reanalysis for a single timestep of an identified East Coast ETC, which was also identified as an AR-concurrent Bomb Cyclone. Precipitation is masked to only include data within the chosen SLP closed loop.

This method was not applied to the entire ETC database, but future work on the project should include this. The source code for this method is available on this project's Github repository in the "Segmenting Storm Precipitation" folder.

Figure 10: Segmented Precipitation for an Example Extratropical Cyclone (Dec. 26, 2010)
This ETC was identified by our previously described methods as an atmospheric river concurrent bomb cyclone. This animation shows the identified SLP loop mask (outlined in red) found for each timestep of a portion of the storm. For each timestep, the precipitation (from ERA-5 reanalysis) was masked with the timestep's SLP loop mask so that only precipitation data within the SLP loop was included. The animation shows this masked precipitation for five timesteps of the storm.

Output: a Database of ETCs

Extratropical Cyclones (ETCs) were identified using the above methods for the U.S. East Coast and West Coast in MERRA-2 and ERA5 reanalysis data and stored in databases used in the analysis described below. The database containing ETCs within MERRA-2 included bomb cyclone identification and atmospheric river concurrency identification based on atmospheric rivers identified in GISS ModelE2.1 nudged to MERRA-2 horizontal winds (LeGrande et al. 2024) due to availability at the time of creating the database. The MERRA-2 database included data from 2007-2021.

The database containing ETCs within ERA5 included bomb cyclone identification but did not include atmospheric river concurrency metrics due to data availability and resolution processing issues. The MERRA-2 database included data from 1950-2023. In the future, the algorithms used to create the ETC databases should be rerun using atmospheric river data in MERRA-2 and ERA5 identified by Guan and Waliser 2024, of which the global database of ARs in both MERRA-2 and ERA5 is available on UCLA Dataverse).

Back to the table of contents

Extratropical Cyclone Identification Analysis

Storm Track Density Plots

Once the database of ETCs was completed, we began by plotting the ETC storm tracks. Because of the large number of ETCs in the dataset, these storm track plots were not very readable past the monthly time scale, so we developed storm track density plots to translate this information. Below is an example of a storm track plot of which the density plots are based on.

Figure 11: Bombogenesis Tracks of Bomb Cyclones in MERRA-2, East Coast USA 2007-2021
The lines represent the path of bombogenesis of each storm where the west most point is the 'start time' of bombogenesis and the east most point is the 'end time' of bombogenesis where the pressure drop threshold was met. The red lines represent Bomb Cyclones that were identified as AR-concurrent and the blue lines represent Bomb Cyclones that were identified as not being AR-concurrent.

Note that the density plots are smoothed using a Gaussian filter (sigma=1) for easier viewing. The code used to create these plots is available in this project's Github repository in the "Storm Track and Density Plots" folder.

ETCs Identified in MERRA-2 Reanalysis (2007-2021)

U.S. East Coast

• 934 Extratropical Cyclones were identified (62 per year on average)
• 884 of those ETCs were concurrent with Atmospheric River (95%)
• 251 Bomb Cyclones were identified (27% of all ETCs, 17 per year on average)
• 212 of those Bomb Cyclones were concurrent with an Atmospheric River (84%)

Figure 12: Density Plot of ETC Storm Tracks identified in MERRA-2, East Coast USA 2007-2021

U.S. West Coast

• 313 Extratropical Cyclones were identified (21 per year on average)
• 229 of those ETCs were concurrent with Atmospheric River (95%)
• 25 Bomb Cyclones were identified (27% of all ETCs, 2 per year on average)
• 16 of those Bomb Cyclones were concurrent with an Atmospheric River (84%)

Figure 13: Density Plot of ETC Storm Tracks identified in MERRA-2, West Coast USA 2007-2021

ETCs Identified in ERA5 Reanalysis

U.S. East Coast

• 2007 to 2021
• 730 Extratropical Cyclones were identified (an average of 49 per year)
• 141 Bomb Cyclones were identified (19% of all ETCs, an average of 9 per year)
• The MCMS algorithm settings used to identify global ETCs in ERA5 was slightly different from the settings used for MERRA-2; this may explain the discrepancy between the two databases.
• 1950 - 2023
• 3723 Extratropical Cyclones were identified (an average of 50 per year)
• 694 Bomb Cyclones were identified (19% of all ETCs, an average of 9 per year)

Figure 14: Density Plot of ETC Storm Tracks identified in ERA5, East Coast USA 1950-2023

Figure 15: Density Plot of Bomb Cyclone Storm Tracks identified in ERA-5, East Coast USA 1950-2023

U.S. West Coast

• 2007 to 2021
• 129 Extratropical Cyclones were identified (an average of 9 per year)
• 4 Bomb Cyclones were identified (3% of all ETCs, an average of 0.3 per year)
• The MCMS algorithm settings used to identify global ETCs in ERA5 was slightly different from the settings used for MERRA-2; this may explain the discrepancy between the two databases.
• 1950 - 2023
• 758 Extratropical Cyclones were identified (an average of 10 per year)
• 20 Bomb Cyclones were identified (3% of all ETCs, an average of 0.3 per year)

Figure 16: Density Plot of ETC Storm Tracks identified in ERA5, West Coast USA 1950-2023

Key Takeaways

Based on ERA5 reanalysis 1950-2023, there were nearly 390% more extratropical cyclones on the U.S. East Coast than on the U.S. West Coast, and it was ~6 times more likely for an East Coast ETC to undergo bombogenesis than a West Coast ETC. Based on MERRA-2 reanalysis 2007-2021, there were nearly 200% more extratropical cyclones on the U.S. East Coast than on the U.S. West Coast. It was ~3.5 times more likely for an East Coast ETC to undergo bombogenesis than a West Coast ETC. Within MERRA-2 reanalysis, 84% of Bomb Cyclones on the East Coast were concurrent with an Atmospheric River. In both ERA5 and MERRA-2 reanalysis, extratropical cyclones tend to occur in the Northern Hemisphere between 30N and 50N latitude.

Back to the table of contents

Atmospheric River-Concurrent Bomb Cyclone Case Study

Figure 17: MERRA-2 sea level pressure output for the northeast U.S. during the December 26-27, 2010 and January 3-4, 2018 bomb cyclones. Figures made using NASA Giovanni.
The MERRA-2 output above depicts a drop in sea level pressure as two separate bomb cyclones impacted the Northeastern U.S. We can see the significant drop in sea level pressure, predominantly along and off the Northeast coast as the cyclones tracked through the region. Note that the MERRA-2 reanalysis depicts that the greatest drop in sea level pressure in both cyclones occurred near or around the 40N-70W latitude, longitude lines. This seems to be a consistently common location for developing cyclones to bomb out as they impact the Northeast.

Figure 18: 500 millibar geopotential heights reanalysis (NOAA) during the December 26, 2010 bomb cyclone
NOAA daily weather maps show a seesaw-like pattern across North America, with far lower heights (trough) associated with the bomb cyclone over the eastern half and far higher heights (ridge) over the western half. This ridge sends cold air southward from Canada/arctic regions, which then clashes with the warmer tropical air mass situated over the southern central U.S. This clashing then leads to the rapid development and intensification of a low pressure system as it tracks Northeast toward and up the east coast. (Husain et al., 2018)

Back to the table of contents

ETC Teleconnections Analysis

We studied several atmospheric teleconnections which have some influence on the Eastern United States, to see if they affect extratropical or bomb cyclone formation frequency.

North Atlantic Oscillation (NAO)

Figure 19: 500mb geopotential height association with NAO+.

Figure 20: NAO association with Extratropical and Bomb Cyclone frequency. Comparisons are made by month (top) and by number of days before formation (bottom).

The source code for oscillation/teleconnection analysis is available in this project's Github repository in the 'teleconnection-analysis' folder.

North Atlantic Oscillation (NAO) tracks the pressure difference between the Azores High and the Icelandic Low in the North Atlantic Ocean; an abnormally large pressure difference indicates an NAO+, and an abnormally low pressure difference indicates an NAO-. NAO- is associated with colder East Coast winters, while NAO+ is associated with milder winters there. (El Niño & Other Oscillations)

NAO was compared with extratropical and bomb cyclone formation rates by month, and up to 28 days before storm formation. NAO- within seven days of storm formation was found to be associated with more extratropical cyclones, and NAO- from three to fourteen days before storm formation was found to be associated with more bomb cyclones. Upon storm formation, NAO- is associated with more extratropical cyclones, but not with more bomb cyclones.

Arctic Oscillation (AO)

Figure 21: AO association with Extratropical and Bomb Cyclone frequency. Comparisons are made by month (top) and by number of days before formation (bottom).

Arctic Oscillation (AO) tracks the pressure difference between the Arctic and the Mid-latitudes, and the strength of the westerly winds surrounding the Arctic. An AO- indicates weak westerly winds, allowing Arctic air to escape equatorward, while an AO+ indicates strong westerly winds, leading to colder temperatures in the Arctic and milder temperatures in the mid-latitudes. (El Niño & Other Oscillations)

AO was compared with extratropical and bomb cyclone formation rates by month, and up to 28 days before storm formation. Between October and March, An AO- is associated with higher extratropical cyclone formation rates from zero to 28 days before storm formation, and is associated with higher bomb cyclone formation rates from three to 28 days before storm formation. Upon storm formation, AO- is associated with more extratropical cyclones, but not with more bomb cyclones.

Note the large degree of noise in the two middle columns. This is because AO has a relatively high standard deviation, close to 1.6. there are relatively few days with an AO amplitude of less than 0.5.

Pacific-North American Oscillation (PNA)

Figure 22: PNA association with Extratropical and Bomb Cyclone frequency. Comparisons are made by month (top) and by number of days before formation (bottom).

Pacific-North American Oscillation (PNA) tracks atmospheric patterns in the North Pacific and over the North American continent. Throughout most of the winter, a PNA+ is associated with a cooler East Coast and a warmer West Coast; the reverse is associated with a -PNA. (El Niño & Other Oscillations)

PNA was compared with extratropical and bomb cyclone formation rates by month, and up to 28 days before storm formation. Between October and March, a PNA+ is associated with higher extratropical and bomb cyclone formation rates from zero to 21 days before storm formation. Upon storm formation, PNA+ is associated with more extratropical and bomb cyclones.

El Niño / La Niña (ENSO)

Figure 23: ENSO association with Extratropical and Bomb Cyclone frequency, by month.

El Niño / La Niña state tracks the sea surface temperature anomaly in the Eastern Equatorial Pacific Ocean; El Niño indicates positive anomalies in this region, while La Niña indicates negative anomalies. This oscillation has an irregular period of two to ten years. (El Niño & Other Oscillations)

Extratropical and bomb cyclone formation rates were also compared to ENSO state across all months. However, no statistically significant difference in ETC or BC formation rates was observed based on ENSO state, in extended winter (ONDJFM) or winter (DJF).

Key Takeaways

Extratropical cyclone frequency is higher during PNA+ and NAO- events, and 3 to 28 days after AO- events, and bomb cyclone frequency is higher during PNA+ events, and 3 to 28 days after AO- events. We observed little measured influence on extratropical or bomb cyclone frequency from ENSO state.

El Niño/La Niña (ENSO) Case Study

Figure 24: Time averaged maps of monthly SLP and satellite precip estimates during 3 La Niña winters (2011-2012), (2016-2017), and (2022-2023) respectively (top to bottom)

The La Niña ENSO state during the winter months is strongly associated with higher mean sea level pressures over the southeast U.S. This area of sinking air unsurprisingly cuts off the subtropical jet and weakens/prevents the development and cyclonic organization of many atmospheric impulses that may move through the region. We can see a lack of significant precipitation near and off of the southeast U.S. coast during the three La Niña winters, with the core of the highest precip amounts falling far to the Northeast out in the open atlantic. This implies that ETC development is delayed until the impulses move further out to sea (away from the influence of the dominant southeast ridge.) It is this delay in intensification that hampers the frequency of bomb cyclones along and off the east coast during La Niña winters.

Figure 25: Time averaged map of monthly SLP and satellite precip estimates during Non-La Niña winter (2009-2010)

Here, a randomly selected Non-La Niña winter shows the characteristics of a far more active subtropical jet. We can see that the mean sea level pressures are significantly lower when compared to the other three La Niña winters. This implies a far less hostile environment for the rapid development of ETCs near and off the southeast coast. It is also evident that the highest precip amounts are located far Southwest relative to the other three La Niña winter case studies, which indicates that ETCs and bomb cyclones develop far quicker, and closer to the coast during non-La Niña winters when compared to La Niña winters.

Back to the table of contents

Research Conclusions and Next Steps

Of the identified extratropical cyclones and the eastern and western coasts of the United States, there were markedly less extratropical cyclones and bomb cyclones on the west coast than the east coast. Extratropical Cyclones on the U.S. East Coast are more likely to occur and experience bombogenesis, and to be concurrent with an atmospheric river, than on the west coast.

We found that within the reanalysis data, extratropical cyclone frequency is higher during PNA+ and NAO- events, and after AO- events, and bomb cyclone frequency is higher during PNA+ events, and after AO- events. We observed little measured influence on extratropical or bomb cyclone frequency from ENSO or AMO state.

Next Steps

• Apply the storm-specific precipitation segmentation algorithm to the ETC datasets.
• Switch the AR-concurrency algorithm to use reanalysis specific AR data rather than data from the GISS model.
• Determine how to apply the ETC identification algorithm used in the MERRA-2 and ERA-5 Fronts ETC data product to the GISS ModelE.
• Determine the GISS model’s accuracy at predicting the frequency and intensity of ETCs, by applying the methods to nudged model runs.
• Apply the methods to future runs of the GISS model under different relative concentration pathways (RCPs) to determine how future climate change scenarios may affect ETC representation and intensity.
• Analyze why observed (and inferred) ETC and bomb cyclone changes are occurring.

Back to the table of contents

Education and Community Outreach

The education component of the CCRI experience centers on creating a unit plan designed to make complex scientific concepts about extreme precipitation accessible and engaging for high school students. The unit plan leverages ArcGIS StoryMaps, allowing students to explore and present data about extreme weather events in their own neighborhoods or hometowns. This approach ensures that learning is relevant and impactful by connecting it to students' personal experiences and communities. Additionally, active involvement in community STEM engagement events fosters a love for science and technology among diverse groups of people including students, teachers, school and district administrators, and broader audience who are interested in STEM education. These events range from workshops and presentations to interactive activities, all aimed at making STEM education accessible and exciting for everyone, regardless of their background. By involving the community in these educational experiences, the goal is to inspire the next generation of scientists and engineers and to demonstrate the practical importance of scientific research in everyday life.

Educator's Unit Plan

• Lesson 1 - Extreme weather in the news: Creating a short news report with data from NASA Worldview
• Lesson 2 - Science behind extreme weather and climate change: Covering background knowledge on extreme weather using NASA Vital Signs
• Lesson 3 - Using data to decode extreme weather: Data analysis activity using ground-based and NASA GPM (global precipitation measurement mission) data via NASA Giovanni (web based visualization)
• Lesson 4 - All about extreme rain: Group project to create ArcGIS Storymap about extreme precipitation
• Capstone - Beyond the Rain: Past, Present, and Future in Our Town: ArcGIS Story Map to increase awareness of the impacts of extreme weather events in their town

Example of Student Capstone Project

Community STEM Engagement Events

• Hosted an Educational Downlink with the International Space Station for Brighton Elementary School in rural Tennessee, taught BES students STEM workshops and managed student question recording (October 2023)
• Lily attended the Space Exploration Educators Conference (SEEC) at Space Center Houston representing the New York Space Grant consortium (February 2024)
• Taught astrobotany workshop at Rensselaer Polytechnic Institute’s Black Families Technology Awareness Day (BFTAD) and gave tours to 6-8th grade students of vertical farming labs (February 2024)
• Over 1,400 students across New York participated in our eclipse science workshops and received free eclipse viewing glasses (April 2024)
• We hosted educational programming at the AI Expo for National Competitiveness in Washington, DC; coordinated local DC Public School students attending the conference; taught Industry 4.0 workshop; organized an AI in Education panel with another CCRI team for the students; and Lily gave a keynote at the conference on the importance of STEM education for national competitiveness (May 2024)
• We coordinated a media partnership with the U.S. Department of Education to film CCRI participant interviews for the Ed. Department’s social media for mentorship month (January 2025) and Sangmin Pak (educator team member) guest wrote for the Ed. Department's blog (April 2024 - August 2024)

Educator’s takeaway from CCRI experience (from Sangmin Pak)

• Participated in cutting-edge science
• Formed valuable connections with experts and fellow educators
• Project impacts were multiplied through turnkeying the new knowledge to my students
• Opportunity to publish Unit Plan, multiplying project impacts to even more classrooms
• Pushed me to be a leader in my school and district!

Back to the table of contents

More about CCRI and Alignment with NASA Objectives

The Climate Change Research Initiative (CCRI) program is an interdisciplinary, collaborative, year-long STEM engagement, and experiential learning opportunity for educators and graduate students to work directly with NASA scientists and lead research teams in a NASA research project hosted at NASA's Goddard Institute for Space Studies; Columbia University's Lamont Doherty Earth Observatory; CUNY City College of Technology in New York City; or NASA's Goddard Space Flight Center in Greenbelt, MD. The summer component of each CCRI project also includes undergraduate and high school interns.

Project Alignment with NASA Objectives:

Back to the table of contents

Project Codebase

The analysis and visualiation code used in this project is open-sourcea and available in the project's Github repository.

Back to the table of contents

References

• Booth, J. F., C. M. Naud, and A. D. Del Genio, 2013: Diagnosing Warm Frontal Cloud Formation in a GCM: A Novel Approach Using Conditional Subsetting. J. Climate, 26, 5827–5845. doi.org/10.1175/JCLI-D-12-00637.1
• Chen, D., Dai, A., & Hall, A. (2021). The convective-to-total precipitation ratio and the “drizzling” bias in climate models. Journal of Geophysical Research: Atmospheres, 126, e2020JD034198. doi.org/10.1029/2020JD034198
• El Niño & Other Oscillations. Woods Hole Oceanographic Institution. (2019, February 6).
• ERA-5: European Centre for Medium-Range Weather Forecasts, Copernicus Climate Change Service, ERA5 hourly data on single levels. doi.org/10.24381/cds.adbb2d47
• GISTEMP Team, 2024: GISS Surface Temperature Analysis (GISTEMP), version 4. NASA Goddard Institute for Space Studies. Most recent publciation: Lenssen, N., G. Schmidt, J. Hansen, M. Menne, A. Persin, R. Ruedy, and D. Zyss, 2019: Improvements in the GISTEMP uncertainty model. J. Geophys. Res. Atmos., 124, no. 12, 6307-6326, doi:10.1029/2018JD029522.
• Global Modeling and Assimilation Office (GMAO)(2015), MERRA-2 tavg1_2d_flx_Nx: 2d,1-Hourly,Time-Averaged,Single-Level,Assimilation,Surface Flux Diagnostics V5.12.4, Greenbelt, MD, USA, Goddard Earth Sciences Data and Information Services Center (GES DISC), 10.5067/7MCPBJ41Y0K6
• GPM IMERG Final Precipitation L3 Half Hourly 0.1 Degree x 0.1 Degree V07. doi: 10.5067/GPM/IMERG/3B-HH/07.
• Kelley, M., G.A. Schmidt, L. Nazarenko, S.E. Bauer, R. Ruedy, G.L. Russell, A.S. Ackerman, I. Aleinov, M. Bauer, R. Bleck, V. Canuto, G. Cesana, Y. Cheng, T.L. Clune, B.I. Cook, C.A. Cruz, A.D. Del Genio, G.S. Elsaesser, G. Faluvegi, N.Y. Kiang, D. Kim, A.A. Lacis, A. Leboissetier, A.N. LeGrande, K.K. Lo, J. Marshall, E.E. Matthews, S. McDermid, K. Mezuman, R.L. Miller, L.T. Murray, V. Oinas, C. Orbe, C. Pérez García-Pando, J.P. Perlwitz, M.J. Puma, D. Rind, A. Romanou, D.T. Shindell, S. Sun, N. Tausnev, K. Tsigaridis, G. Tselioudis, E. Weng, J. Wu, and M.-S. Yao, 2020: GISS-E2.1: Configurations and climatology. J. Adv. Model. Earth Syst., 12, no. 8, e2019MS002025, doi:10.1029/2019MS002025.
• LeGrande, A. N., Booth, J. F., Naud, C. M., Ordaz, C., & Crespo, J. A. (2024). Just how river-like are atmospheric rivers? Geophysical Research Letters, 51, e2023GL105828. doi.org/10.1029/2023GL105828
• Martinkova, M., and J. Kysely, 2020: Overview of observed Clausius-Clapeyron scaling of extreme precipitation in Midlatitudes. Atmosphere, 11, 786, https://doi.org/10.3390/atmos11080786
• Myhre, G., Alterskjær, K., Stjern, C.W. et al. Frequency of extreme precipitation increases extensively with event rareness under global warming. Sci Rep 9, 16063 (2019). doi.org/10.1038/s41598-019-52277-4
• Naud, Catherine M. Clouds and Precipitation in Extratropical Cyclones: A Global Satellite Climatology for GCM Evaluation. Studies of Cloud, Convection and Precipitation Processes Using Satellite Observations. December 2022, 289-313. doi.org/10.1142/9789811256912_0015
• Sanders, F.; Gyakum, J.R. Synoptic-dynamic climatology of the “bomb”. Mon. Weather Rev. 1980, 108, 1589–1606. doi.org/10.1175/1520-0493(1980)108<1589:SDCOT>2.0.CO;2
• Seiler, C., Zwiers, F.W. How well do CMIP5 climate models reproduce explosive cyclones in the extratropics of the Northern Hemisphere?. Clim Dyn 46, 1241–1256 (2016). doi.org/10.1007/s00382-015-2642-x
• Zhang, Z., Ralph, F. M., & Zheng, M. (2019). The relationship between extratropical cyclone strength and atmospheric river intensity and position. Geophysical Research Letters, 46. doi.org/10.1029/2018GL079071
• Zhang, L., Zhao, Y., Cheng, T. F., & Lu, M. (2024). Future changes in global atmospheric rivers projected by CMIP6 models. Journal of Geophysical Research: Atmospheres, 129, e2023JD039359. doi.org/10.1029/2023JD039359

  Back to the table of contents