Department of Energy - Critical Review of Impacts of GHG Emissions on the US Climate July 2025

The hemispheric symmetry of the albedo is a simple gross metric for climate models. Rugenstein and Hakuba (2023) defined that metric as the difference between the NH and SH annual mean albedos, expressed as Wm−2 of the reflected sunlight, and compiled it for the CMIP6 climate models, as shown in Figure 5.8. Most of the CMIP6 models do not reproduce the small observed asymmetry (about 0.1 Wm-2) and even disagree as to which hemisphere is more reflective. Moreover, the magnitude of the asymmetry ranges up to 5 Wm-2 in some models, twice as large as the current anthropogenic forcing (about 2.7 Wm-2). The significance of unphysical albedo asymmetries in the climate models is not yet fully known. However, other model studies suggest that interhemispheric changes in albedo can alter poleward heat fluxes, meridional temperature gradients, storminess, and differences in hemispheric ocean heat storage. The discrepancy between models and observations further raises issues regarding cloud feedback processes and so more generally diminishes confidence in model projections of the future climate.

41

Figure 5.8. Differences in 20-year average reflectivity (albedo) between the Northern and Southern Hemispheres for CMIP6 models used in the most recent IPCC assessment (colored bars). The very small observed difference is indicated by the vertical black line. From Rugenstein and Hakuba (2023).

5.8

U.S. Corn Belt

One of the largest discrepancies between models and observations is in the U.S. Corn Belt, a region of particular importance to global food production. Figure 5.9 shows the warming trends for summertime (June, July, August) for the 12-state Corn Belt (IN, IA, IL, ND, SD, MO, MN, WI, MI, OH, KS, NE) during 1973-2022. All 36 climate models (red) warm far too rapidly compared to observations (blue).

42

Figure 5.9: Modeled versus observed warming trends in the U.S. Corn Belt, 1973-2022.

As discussed in Chapter 9, the anticipated negative effects of increasing temperatures on U.S. corn yields have not materialized, in contrast to widely publicized studies proclaiming that theoretical future impacts are already being experienced (e.g., Seager et al., 2018). The IPCC acknowledges limitations in the accuracy of regional climate model outputs. This example shows that users need to assess model projections carefully on a case-by-case basis since local biases might be sufficiently large that the models are simply not fit for purpose. As has recently been noted by two leaders of the modeling community (emphasis added) … for many key applications that require regional climate model output or for assessing largescale changes from small scale processes, we believe that the current generation of models is not fit for purpose. (Palmer and Stevens 2019) To summarize: • •

Climate models show warming biases in many aspects of their reproduction of the past few decades. They produce too much warming at the surface (except in the models with lowest ECS), too much warming in the lower-and mid-troposphere and too much amplification of warming aloft 43

• • •

They also produce too much stratospheric cooling, too much snow loss, and too much warming in the U.S. Corn Belt. The hemispheric albedo difference in individual climate models ranges widely in sign and magnitude compared to observations. The range in W/m2 is three times larger than the direct anthropogenic forcing of CO2. The IPCC has acknowledged some of these issues but not others.

References Christy, J. R., R. T. McNider (2017). Satellite bulk tropospheric temperatures as a metric for climate sensitivity. Asia-Pacific Journal of Atmospheric Sciences 53, 511-518. https://doi.org/10.1007/s13143-017-0070-z Christy, J. R. (2022). Time series construction of Oregon and Washington snowfall since 1890 and an update of California snowfall through 2020. Journal of Hydrometeorology 23, 1845-1860. https://doi.org/10.1175/JHM-D-21-0178.1 Connolly, R., M. Connolly, W. Soon, et al. (2019). Northern Hemisphere snow-cover trends (1967– 2018): A comparison between climate models and observations. Geosciences 9, 135. https://doi.org/10.3390/geosciences9030135 Datseris, G., B. Stevens (2021). Earth’s albedo and its symmetry. AGU Advances 2, e2021AV000440. https://doi.org/10.1029/2021AV000440 Karl, T.R., S. J. Hassol, C. D. Miller, W. L. Murray, eds. (2006). Temperature Trends in the Lower Atmosphere: Steps for Understanding and Reconciling Differences. U.S. Climate Change Science Program/Subcommittee on Global Change Research. https://digital.library.unt.edu/ark:/67531/metadc12017/m2/1/high_res_d/sap1-1-final-all.pdf Klotzbach, P.J., R. A. Pielke, Sr., R. A. Pielke, Jr., J. R. Christy, R. T. McNider (2009). An alternative explanation for differential temperature trends at the surface and in the lower troposphere. Journal of Geophysical Research — Atmospheres 114(D21).

https://doi.org/10.1029/2009JD011841 McKitrick, R. and J. R. Christy (2020). Pervasive warming bias in CMIP6 tropospheric layers. Earth and Space Science 7. https://doi.org/10.1029/2020EA001281 McKitrick, R., and J.R. Christy (2025), “Data and Code for DoE Report 2025”, Mendeley Data, V1, doi: 10.17632/8n9ks93vn7.1 Palmer, T., and B. Stevens (2019). The scientific challenge of understanding and estimating climate change. Proceedings of the National Academy of Sciences 116(49), 24390–24395. https://doi.org/10.1073/pnas.1906691116 Philipona, R., C. Mears, M. Fujiwara, et al. (2018). Radiosondes show that after decades of cooling, the lower stratosphere is now warming. Journal of Geophysical Research—Atmospheres 123. https://doi.org/10.1029/2018JD028901 Rugenstein, M., and M. Hakuba (2023). Connecting hemispheric asymmetries of planetary albedo and surface temperature. Geophysical Research Letters 50, e2022GL101802. https://doi.org/10.1029/2022GL101802 Santer B D, P.W. Thorne, L Haimberger et al. (2008) Consistency of modelled and observed temperature trends in the tropical troposphere International Journal of Climatology 28 1703–22 https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/joc.1756 44

Santer, B.D. S. Po-Chedley, L. Zhao, C et al. (2023) Exceptional stratospheric contribution to human fingerprints on atmospheric temperature, Proceedings of the National Academy of Sciences U.S.A. 120 (20) e2300758120, https://doi.org/10.1073/pnas.2300758120. Scaffeta, N (2023) CMIP6 GCM ensemble members versus global surface temperatures. Climate Dynamics 60, 3091–3120 (2023). https://doi.org/10.1007/s00382-022-06493-w Seager, R., J. Feldman, N. Lis, et al. (2017). Whither the 100th meridian? The once and future physical and human geography of America’s arid–humid divide. Part II: The meridian moves east. Earth Interactions 22. https://doi.org/10.1175/EI-D-17-0012.1 Spencer, R. W. (2024). Global warming: Observations vs. climate models. Environment Backgrounder, The Heritage Foundation. https://www.heritage.org/environment/report/global-warming-observations-vs-climate-models Stephens, G. L., D. O'Brien, P.J. Webster et al. (2015). The albedo of Earth. Reviews of Geophysics. https://doi.org/10.1002/2014RG000449 Vogelsang, T. and N. Nawaz (2016). Estimation and inference of linear trend slope ratios with an application to global temperature data: Journal of Time Series Analysis 38. https://doi.org/10.1111/jtsa.12209

45

6 EXTREME WEATHER Chapter Summary Most types of extreme weather exhibit no statistically significant long-term trends over the available historical record. While there has been an increase in hot days in the U.S. since the 1950s, a point emphasized by AR6, numbers are still low relative to the 1920s and 1930s. Extreme convective storms, hurricanes, tornadoes, floods and droughts exhibit considerable natural variability, but long-term increases are not detected. Some increases in extreme precipitation events can be detected in some regions over short intervals but the trends do not persist over long periods and at the regional scale. Wildfires are not more common in the U.S. than they were in the 1980s. Burned area increased from the 1960s to the early 2000’s, however it is low compared to the estimated natural baseline level. U.S. wildfire activity is strongly affected by forest management practices.

6.1

Introduction

High impact weather extremes, usually related to temperature, precipitation and/or high wind, can disrupt infrastructure and therefore endanger human health and wellbeing. The issue is not whether extremes occur. Rather, it is whether there are long-term (decadal scale) changes in the frequency or character of extremes (“detection”), as well as the extent to which such changes and the attendant changes in hazards are caused by anthropogenic emissions of greenhouse gases (“attribution”; e.g., AR6 Seneviratne et al. 2021). Process-based understanding and simple thermodynamic arguments have been invoked to assert that warming is worsening extreme weather events. However, it is naïve to assume that any recent extreme event is caused by human influences on the climate. Climate is about the statistical properties of weather over decades, not single events. Further, there are only about 130 years of reliable observational records that can be analyzed statistically. That brief interval does not begin to contain all the extreme events that the climate system can create on its own. Over geologic time the climate system has generated an (essentially) infinite variety of weather patterns and extremes that humans have never observed and thus are absent from the databases used to determine extreme statistics [see Perils of short data records below].

For that reason, attributing an extreme event unprecedented in the record requires assumptions about the magnitude of natural variations. This chapter is concerned with detection of trends in extreme weather, while Chapter 8 considers causal attribution, with Section 8.4 specifically addressing extreme weather. If no trend is detected, then clearly there is no basis for attribution. But even where a trend is observed, attribution to human-caused warming does not necessarily follow. This is especially true for precipitation events. The hydrological literature has long noted the presence of long, slow and irregular oscillations in rainfall data (Hurst 1951, Cohn and Lins 2005, Markonis and Koutsoyiannis 2016). The characteristics of these natural patterns require long records to accurately estimate variability. Analysis of records that are short relative to the scale of natural variability will tend to misrepresent trends, therefore overstating the significance of apparent trends and underestimating the likelihood of extreme events (Cohn and Lins 2005). 46

Figure 6.1.1: The annual minimum depth of the Nile River near Cairo over more than 650 years from 622 to 1284 A.D. The data, measured in meters, shows a characteristic pattern of year-to-year fluctuations around longer-term trends. Data from Koutsoyiannis (2013)

A good example of this is the eight-century long record of the annual minimum height of the Nile River observed at Roda Island in Cairo shown in Figure 6.1.1. The Nile River is fed by precipitation over a 4 million square mile drainage basin, an area equal to about one third of CONUS. Since human influences on the global climate were negligible long before the 20th century, the century-scale variability of the thirtyyear average is entirely natural; Egyptians of the seventh and eight centuries would have been incorrect to assume that the worsening drought during that time was the “new normal.” With these caveats in mind, we examine the evidence for changes in selected weather and climate extremes. A recurring theme is the wide gap between public perceptions and scientific evidence. It has become routine in media coverage, government and private sector discussions, and even in some academic literature to make generalized assertions that extreme weather of all types is getting worse due to GHGs and “climate change.” Yet expert assessments typically have not drawn such sweeping conclusions and instead have emphasized the difficulty both of identifying specific trends and establishing a causal connection with anthropogenic forcing.

In the sections to follow we provide excerpts from various IPCC and NCA assessment reports denoting the sources as follows: SREX: The IPCC Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (2012) AR6: The IPCC Sixth Assessment Report Working Group 1 (2021). NCA4: The U.S. Climate Science Special Report of the Fourth National Climate Assessment (2017) Volume I. NCA5: Fifth National Climate Assessment (2023). In the excerpts. italics are in the original whereas boldface emphasis has been added. 47

Additionally we use standard government sources to provide information through 2024 wherever possible.

6.2

Hurricanes and tropical cyclones

AR6 provides the following assessment of tropical cyclones (TCs; used here as a synonym for hurricanes): AR6: There is low confidence in most reported long-term (multidecadal to centennial) trends in TC frequency or intensity-based metrics due to changes in the technology used to collect the best-track data. (IPCC, 2021 p. 1585) AR6: It is likely that the global proportion of major (Category 3–5) tropical cyclone occurrence has increased over the last four decades . . . There is low confidence in long-term (multi-decadal to centennial) trends in the frequency of all-category tropical cyclones. (IPCC, 2023 SPM p. 9) AR6: A subset of the best-track data corresponding to hurricanes that have directly impacted the United States since 1900 is considered to be reliable, and shows no trend in the frequency of U.S. landfall events. (IPCC 2021 p. 1585) Since 1980, when satellite observations first fully covered the global oceans, we have confidence in the numbers of total global hurricanes and major hurricanes (Category 3 and higher). Figure 6.2.1 shows that on average, each year there are about 50 hurricanes, with about 25 reaching major hurricane status (Maue, 2025). There is substantial year-to-year and decadal variability, a weak decrease in the number of hurricanes, and a slight but insignificant increase in the number of major hurricanes.

These two trends combine to create an overall increase in the proportion of major hurricanes. Global hurricane statistics are dominated by the Northwest Pacific Ocean, which accounts for ~35 percent of total global hurricanes, whereas the Atlantic accounts for ~15 percent of global hurricanes (Colorado State University, 2025). Data in the Atlantic basin extends further back than in the other ocean basins and is most relevant to U.S. policy makers.

48

Figure 6.2.1: Global frequency of hurricanes and major hurricanes since 1980. Source: Updated from Maue 2011.

Figure 6.2.2 shows the frequency of Atlantic hurricanes and major hurricanes (Category 3 and higher) back to 1920. Data prior to 1965 (the onset of satellite observations in the Atlantic, shaded in blue) shows some undercounting, with data prior to 1920 showing substantial undercounting (Vecchi and Knutson, 2011). All measures of Atlantic hurricane activity show a significant increase since 1970. However, the period from 1971-1994 saw exceptionally low activity, with high activity (comparable to the past two decades, even with undercounting) also observed during the 1950’s and 1960’s, and even in the 1930’s.

49

Figure 6.2.2: Atlantic frequency of hurricanes (HR) and major hurricanes (MHR) since 1920. Source National Hurricane Center (2024)

Figure 6.2.2 shows that Atlantic hurricanes vary strongly on decadal and multidecadal time scales. These variations are associated primarily with the Atlantic Multidecadal Oscillation (AMO), which is manifest in basin-wide sea surface temperature and sea level pressure fluctuations connected to large scale ocean circulation patterns. The AMO was in its warm phase during 1926-1970 and 1995-present, but in its cool phase during 1971-1994. It has its greatest impact on the number of major hurricanes (Category 3+), which Goldenberg et al. (2001) associated with above normal SSTs and decreased vertical shear in the AMO warm phase (see also Bell and Chelliah, 2006; Klotzbach et al., 2018). Klotzbach et al. (2018) conducted a comprehensive evaluation of the landfalling hurricane data for the Continental U.S. since 1900. Figure 6.2.3 updates their analysis through 2024. While the largest numbers of landfalling hurricanes are from 2004, 2005 and 2020, there is no statistically significant trend since 1920. Figure 6.2.3 also shows the time series for major hurricane landfalls (Category 3-5). The largest year in the record is 2005, with 4 major hurricane landfalls. However, following 2005 there were no major hurricanes striking the U.S through 2016, the longest such period since 1920.

50

Figure 6.2.3: U.S. landfalling frequency of hurricanes (HR) and major hurricanes (MHR) since 1920. Source NOAA HRD(a) (2024)

Figure 6.2.3 shows substantial interannual to multidecadal variability in U.S. landfall activity. Klotzbach et al. (2018) examined how the landfall counts vary with ENSO (El Niño versus La Niña) and the warm versus cold phases of the Atlantic Multidecadal Oscillation (AMO). Villarini et al. (2012) provide an analysis of U.S. hurricane landfalls back to 1878. While it is possible that some landfalls were missed in the late 19th century owing to sparsely populated regions on the Gulf Coast, it is remarkable that the highest year in the entire record is 1886, with 7 hurricane landfalls, when human influences on the climate were much smaller than they are today. Table 6.2.1 shows the 10 strongest hurricanes (plus ties) to make U.S. landfall. Of the hurricanes that have made landfall with sustained winds greater than 150 mph, only one has occurred in the 21st century. In summary, analyses of both global and regional variability and trends of hurricane activity provide the basis for detecting changes and understanding their causes. The relatively short historical record of hurricane activity, and the even shorter record from the satellite era, is not sufficient to assess whether recent hurricane activity is unusual relative to the background natural variability.

Atlantic hurricane processes are influenced substantially by the natural modes of ocean circulation variability in the Atlantic, notably the Atlantic Multidecadal Oscillation. While it has long been hypothesized that a rising global sea surface temperature would cause an increase in hurricane intensity, identification of any significant trend in the hurricane data is hampered by a short data record and substantial natural variability. Process-based understanding also suggests that storm surges and rainfall from hurricanes should be increasing with warmer temperatures. However, the relatively small number of hurricanes with varying landfall locations and the complex dynamics associated with each storm preclude meaningful detection of change. 51

Rank

Year

1 2 3 4 5 5 5 5 5 5 5 5

1935 1969 1992 2018 1856 1886 1919 1932 2004 2020 2021 2022

Landfall Wind (MPH) 185 175 165 160 150 150 150 150 150 150 150 150

Name "Labor Day" Camille Andrew Michael "Last Island" "Indianola" --------"Freeport" Charley Laura Ida Ian

Table 6.2.1 Strongest hurricanes to make landfall along the U.S. coast. Source (NOAA HRD(b), 2024)

6.3

Temperature extremes

The AR6 assessment focused on the period after 1950 and reported increasing trends in heatwave frequency and intensity. However, NCA4 noted that heatwave activity in the U.S. reached a peak in the 1930s (Figure 6.3.1). AR6: It is virtually certain that hot extremes (including heatwaves) have become more frequent and more intense across most land regions since the 1950s, while cold extremes (including cold waves) have become less frequent and less severe (SPM, A3.1) AR6: In North America, there is very robust evidence for a very likely increase in the intensity and frequency of hot extremes and decrease in the intensity and frequency of cold extremes for the whole continent, though there are substantial spatial and seasonal variations in the trends. Minimum temperatures display warming consistently across the continent, while there are more contrasting trends in the annual maximum daily temperatures in parts of the USA. (Chapter 11, p 1550) NCA4: Changes in warm extremes are more nuanced than changes in cold extremes. For instance, the warmest daily temperature of the year increased in some parts of the West over the past century, but there were decreases in almost all locations east of the Rocky Mountains. In fact, all eastern regions experienced a net decrease, most notably the Midwest (about 2.2°F [1.2°C]) and the Southeast (roughly 1.5°F [0.8°C]). (pp.

190-191) NCA4: Since the mid-1960s, there has been only a very slight increase in the warmest daily temperature of the year (amidst large interannual variability). Heat waves (6-day periods with a maximum temperature above the 90th percentile for 1961–1990) increased in frequency until the mid-1930s, became considerably less common through the mid-1960s, and increased in frequency again thereafter. As with warm daily temperatures, heat wave magnitude reached a maximum in the 1930s. (pp. 190-191)

52

Figure 6.3.1: U.S. Heat waves since 1900. Source: NCA4 Figure 6.4

6.3.1 Temperatures in the U.S. are becoming less extreme Daily maximum temperatures in the warm season (Tmax, May-Sep) and daily minimum temperatures in the cold season (Tmin, Dec-Mar) are available beginning in Dec 1898 (126 years). The dataset consists of 1,211 CONUS stations designated as United States Historical Climate Network or USHCN stations (see Figure 6.3.2; Quinlan et al. 1987, Karl et al. 1990). These stations were selected by NOAA as having the fewest problematic issues with gaps, station moves, and instrument changes. Where gaps still exist, nearby stations (bias-corrected) were merged so that the median volume of data available for a station is 98%. Although there are certainly errors in the dataset, including unresolved spurious warming due to UHI effects that especially bias Tmin records, this data set is sufficiently accurate for assessing trends in Tmax heat extremes.

53

Figure 6.3.2 Locations of USHCN temperature stations. Source: USHCN.

We begin with the question of whether the occurrence of daily record high or low temperatures has changed since Dec 1898. Each warm-season has 153 days (1 May to 30 Sep) and each cold-season has 122 days (1 Dec to 31 Mar). For each station and day, we calculated the year in which the record highest (lowest) temperature occurred. With 126 years of observations, if there were no temperature trends over time, the expected number of records for Tmax would be 1.21 (=153/126) per station per year and for Tmin 0.96 (=122/126) per station per year. Figure 6.3.3 shows the observed distribution in time of the occurrence of these extreme events. There is a common feature in many metrics of warm-season extremes in the CONUS - the exceptional heat of the 1920s and especially the 1930s, peaking in 1936. On a per-station average, 60 percent of the Tmax records and 59 percent of the Tmin records occurred in the first half of the period (1899-1961).

54

Figure 6.3.3 Number of daily record High (red) and Low (blue) temperatures for warm and cold seasons in the CONUS. The lines represent the 15-year running, centered average. Source: 1,211 USHCN stations supplemented as needed to achieve a minimum of 92 percent of observations in the 126-year period since Dec 1898. US48: contiguous U.S. states. Tmax: maximum temperature. Tmin: minimum temperature.

On the cold side, the Valentine’s Day Arctic outbreak in Feb 1899 stands as the most extensive cold extreme experienced by CONUS, with 1917 in 2nd place. The frequency of cold records has declined, especially over the last quarter of the period in which only 13 percent of the extreme cold events were measured. In contrast, 25 percent of the extreme Tmax records were achieved in the last quarter, in accordance with statistical expectations. These general features have been noted in past assessments (see above, IPCC AR6, NCA4). Combining the two histories, the overall reduction in numbers of both cold and hot extremes over the past century indicates a climate less prone to extremes. This pattern is also shown in Figure 6.3.4. For each station and for each year the hottest warm season and coldest cold-season temperatures were calculated. Then the differences between these were computed by station and geographically-averaged over all stations, thus yielding an annual measure of the expected range of local extreme temperatures for each year. Figure 6.3.4 shows the 15-year trailing average of this measure, which has clearly declined over the past century.

55

Figure 6.3.4. 15-year trailing average of the difference each year of each station’s hottest warmseason Tmax and coldest cold-season Tmin relative to the long-term average. Source: Author analysis of USHCN data.

The average difference for each station between the hottest summer Tmax and coldest winter Tmin has declined by about 5°F in the past 126 years. The decline is due mostly to warmer winter Tmin, but a decline in summer Tmax is also a factor. The rise in Tmin has been strongly related to the growing presence of manufactured surfaces around the weather stations over the last 100+ years (the so-called urban heat island effect; Section 3.3, Karl et al. 1988, Runnals and Oke 2006, and Spencer et al. 2025). In summary, while temperature extremes are regularly experienced in the U.S. and attract a great deal of media attention, long term records show the U.S. climate has become less extreme over time (milder) when measured by the range between warm season maxima and cold season minima. 6.3.2 Exceedances of a heat threshold Under the heading of “The Risk of Temperature Extremes is Changing”, the most recent U.S. National Climate Assessment report (NCA5) notes the increase in a threshold metric of number of days at or above 95°F, stating, The western U.S. has been particularly affected by extreme heat since the 1980s …, experiencing a larger increase in days over 95°F, as would be expected given the greater warming in that region relative to the eastern US. Several major heatwaves have affected the U.S. since 2018, including a record-shattering event in the Pacific Northwest in 2021.

Are the occurrences of 95°F days changing? In a climate as varied as that of CONUS, threshold statistics can be misleading. A region with many stations that have near 95°F Tmax average temperatures in the summer might see large swings in the metric when only small changes in average temperature occur. Elsewhere with stations that either rarely or virtually always achieve 95°F Tmax temperatures, a small change will not have much impact on the results.

56

Figure 6.3.5. Total number of days ≥ 95°F in 6-yr periods, U.S. 48 (bars) and regions (lines). A 6year period is used as this evenly divides the 126-year record. Results are robust to using periods from 2 to 11 days. US48: contiguous U.S. states. See Figure 6.3.2 for region names. Source: Author analysis of USHCN data.

In the past 126 years, the average CONUS station experienced 129 days exceeding 95°C per 6-yr period, but the regional values range from 278 in the Southern Plains to 9 in the Northeast. Thus, such threshold analyses must be interpreted with caution. Figure 6.3.5 shows that only three of the nine regions, all in the West, have experienced upward trends in the number of 95°F or hotter days (dashed lines). The CONUS as a whole has not, and the other six regions have experienced declines. The Pacific NW heatwave of 2021 referenced in the NCA5 quote will be examined more closely in Section 8.6.1. The evidence indicates that it was a single, unprecedented event in the record, not part of a pattern of increasing extreme heat. For example, the 5-day average tropospheric grid-point temperature anomaly over the Pacific NW during that event was +10.8 C, the most extreme Northern Hemisphere grid point summer anomaly in the 46 years from over 4 million grid values. In contrast, the global temperature anomaly during that time was virtually zero (+0.03 C, Mass et al. 2024). 6.3.3 Heatwaves Heatwaves (consecutive days that exceed an extreme threshold) have a greater societal impact than a single daily record temperature.

We measure “Heatwave Days” here as the count of all days in May-Sep each year that exceed the 90th percentile for that day and that lie within a period of at least six consecutive days. This is equivalent to the method used in NCA4, except that the reference period here is the entire record 1899-2024 while NCA4 truncated the reference period to 1961-1990, which was a cool interval. (The pattern of results shown below does not depend on the choice of reference period.) That truncation boosts positive results (days exceeding the 90th percentile) in years warmer than the reference period, especially starting in 1960 and moving to the present (see Figure 6.3.6 and discussion below). 57

Figure 6.3.6 15-year trailing average of number of heatwave days per year per station in the CONUS (black line) and two regions: West (red), Central-east (green).

Figure 6.3.6 indicates that there are regional variations in heatwave activity. The excessive heat of the first half of the 20th century occurred primarily in the eastern two-thirds of the nation, while the West has seen a recent increase of heatwave days (NCA5). This indicates that the background warm season circulation favored heatwaves in the eastern portions of the country in the first half of the 20 th century, but in the 21st century the patterns have favored heatwaves in the West. For CONUS as a whole, heatwaves are no more common today than they were a century ago, consistent with the upper panel of our Figure 6.4.1 taken from NCA4. This metric varies significantly with region. The four northern regions (Pacific NW, Northern Plains, Upper Midwest, and Northeast) on average experience 15 to 27 heatwave days per 15-year period. In contrast, the five southern regions (Pacific Southwest, 4-Corners, Southern Plains, Ohio Valley and Southeast) see 37 to 54 such days, essentially twice as many. This suggests the summer circulation pattern is more prone to stationary events in the southern regions while transient systems in the northern regions are more common and thus cut short these potentially longer events. The analysis of heatwaves is an example of why it is important to consider complete datasets and appropriate metrics.

The NCA5 directs readers to the website https://www.globalchange.gov/indicators/heat-waves (USGCRP 2023) to view a figure showing the number of urban heatwaves by decade from the 1960s, which we reproduce as Figure 6.3.7.

58

Figure 6.3.7: Average number of urban heatwaves per year for 50 large U.S. metropolitan areas, a misleading metric for reasons explained in the text. From https://www.globalchange.gov/indicators/heat-waves (accessed May 22, 2025).

The figure shows a monotonic increase in each decade from two occurrences per year in the 1960s to six in the 2020s. The definition of a heatwave used is an unusual but practical measure of human discomfort - a period of at least 2 consecutive days when the minimum apparent temperature (combination of temperature and humidity) exceeds the 85th percentile. Note too, the dataset is limited to the 50 largest U.S. cities. Given the unusual heatwave definition and urban focus, these increasing values since 1960 presented in USGCRP (2023) are not informative about long term trends or the influence of GHG emissions for at least two reasons. First, as shown in Figures 6.3.1 and 6.3.6, the 1960s was the coldest decade and 1970s the second coldest decade since the 1910’s, so this starting date preconditions the time series to show increases. Second, post-1960 urbanization in these cities is a major factor in the rise of Tmin relative to colocated Tmax and relative to Tmin at nearby rural stations (Karl et al. 1988, Runnals and Oke 2006, Christy et al. 2009, McNider et al.2012). This does not dismiss the real rise in nighttime temperatures in major U. S. cities and the societal impacts associated with these changes.

However, we note for a variety of reasons that summer Tmax (especially in rural areas) is a better metric for detecting changes in heatwaves influenced by changes in the background climate due, for example, to increasing GHGs (Christy et al. 2009). For CONUS as a whole, the evidence in this section suggests GHG emissions have had little-to-no effect on heatwaves against the background of urbanization and natural climate variability. Irrespective of the ultimate cause of regional trends, heatwaves have important effects on society that must be addressed, as we discuss in Chapter 10.

59

BOX: Perils of short data records

San Francisco provides a good case study of the limitation of using short historical samples to characterize natural variability of extreme events. Suppose we use a 130-year sample of daily San Francisco precipitation from 1895 to 2024 and we look for 3-day, 5-day, 14-day and 30-day rainfall records. The results are as shown in Table 6.2. Event

Record Year (inches) 3-day 6.94 2023 5-day 8.55 2023 14-day 12.62 2023 30-day 18.93 1998 Table 6.2: Extreme rainfall records, San Francisco, 1895-2024. The records all cluster in the more recent years. 2023 appears to be an exceptional year and since it is near the end of the sample, it might suggest that the climate has shifted into a more hazardous state, perhaps because of human influences. [Note “2023” indicates the event occurred in the wateryear of Aug 2022 to Jul 2023.] But the picture is very different if we use a sample that begins 45 years earlier, in 1850. Table 6.3 shows that the record-setting events all happened in the 1860s. Furthermore 2023 is now not even in 2nd place but falls to 3rd or 4th place. And comparing the records of the two charts shows the extreme precipitation events in 1862 and 1867 involved considerably more rainfall than the 1998 and 2023 events, with 14- and 30-day totals about 50 percent higher. Rank of 130-yr Record Year extreme listed (inches) in Table 6.2 3-day 8.85 1867 3 5-day 9.80 1867 3 14-day 19.05 1862 4 30-day 28.25 1862 2 Table 6.3: Extreme rainfall records, San Francisco, 1850-2024. Event

The range of natural variability is made even more remarkable when paleoclimate evidence is examined. Porter et al. (2011) discovered that in the past 1,800 years at least six megastorms were more intense than the devastating 1861-62 ARkStorm that struck the region. Evidently such extreme events, “unprecedented” in our 1895-2024 sample, have impacted the region about every 300 years, though not since 1895. This example illustrates the limitations of using relatively short climate periods (~130 years) to assess the character and range of natural variability in general and of extreme events in particular. Accurate representation of the full range of natural variability is necessary for any attribution analyses (Section 8.6). Infrastructure planners, emergency management institutions and attribution scientists would understand the significant mischaracterization of the magnitude of a future extreme if based only on the last 130 years. In this case, a single time-sample of 130 years provides an underestimation of the extreme value by up to 50 percent determined when adding only 45 more years of observations. Compared to millennial-scale paleoclimate evidence, an even greater underestimation would occur. An important lesson is that the climate can deliver great surprises on its own, even without human influences.

60

6.4

Extreme precipitation AR6 assessed that an increase in heavy precipitation has been observed in data starting in the 1950s. AR6: The frequency and intensity of heavy precipitation events have increased since the 1950s over most land area for which observational data are sufficient for trend analysis (high confidence). (SPM A3.2) AR6: In North America, there is robust evidence that the magnitude and intensity of extreme precipitation has very likely increased since the 1950s. Both [one-day maxima] and [5-day maxima] have significantly increased in North America during 1950-2018. (Chapter 11, p. 1560)

The U.S. National Climate Assessments (NCA4, NCA5) have highlighted an increase in the occurrence of the heaviest precipitation events (defined in different ways) primarily in the eastern half of CONUS, especially the Northeast, when starting the analysis in either 1901 or 1958. Interestingly, the regional variations indicate that the largest increases in extreme precipitation events are in the Northeast and the smallest in the West, a pattern counter to the changes in temperature extremes (Figure 6.3.6). McKitrick and Christy (2019) examined long-term and consistent station observations of extreme daily precipitation to test some of these NCA claims for the Southeast and West Coast using a trend model with a non-parametric variance estimator robust to the complex autocorrelation properties of precipitation data. When the time series were extended back in time (as far as 1872 in some cases) or started later (1978), there were no significant trends for either region. These findings have been updated for this report (McKitrick and Christy 2025) with similarly constructed observations from 29 stations on the CONUS Pacific Coast (1893ff from San Diego CA to Blaine WA) and 24 stations in the humid Southeast (1872ff from Austin TX to Washington DC), also adding 27 stations in the Northeast (1888ff from Buffalo NY to Eastport ME). The locations are shown in Figure 6.4.1.

The stations were selected based on availability of long-term high-quality records. The regions are each associated with important features of extreme precipitation behavior: the Pacific coast is associated with landfalling atmospheric rivers for which AR6 cites evidence of increasing activity since 1948 with further increases expected as the world warms (AR6 8.3.2.8.2); the NCA report indicates that the Northeast has experienced the greatest increase in extreme events, and the Southeast is also a place noted in the NCAs as having increased extreme events. The results of applying the analysis of McKitrick and Christy (2019) were as follows for each region, then followed by further explanation. Pacific Coast heavy rainfall events •

The average precipitation trend is statistically significant (downwards) in Astoria OR; insignificant elsewhere.

•

The trend in rainfall variance is positive and significant in Big Sur CA; insignificant elsewhere.

•

The trend in daily maximum precipitation is positive and significant in Aberdeen WA and Big Sur CA and negative and significant in Newport OR (insignificant elsewhere). 61

•

Averaged over all stations in the region, none of these three trend parameters is statistically significant.

The Pacific Coast receives considerable precipitation from Atmospheric River (AR) events which often last more than a day or two (e.g., Gershunov et al. 2017, Pan et al. 2024). The worst series of such events in recent history was the so-called ARkStorm that occurred during December 1861 and January 1862; it dumped nearly 10 feet of rain in parts of California and submerged the entire Central Valley for weeks under as much as 15 feet of water (Brewer 1930, Null and Hulbert 2007). Additionally, paleoclimate research has found six megastorms more severe than 1861–1862 in California during the last 1800 years, occurring at intervals of 300 years or so (Porter et al. 2011).

Figure 6.4.1: Locations of precipitation monitoring stations used in this report. Orange: Pacific coast. Blue: Northeast. Green: Southeast. Data from McKitrick and Christy (2025).

We examine occurrences of 5-day deluges as follows. Taking the Pacific coast as an example, a 130year span contains 26 5-year intervals. At each location we computed the 5-day precipitation totals throughout the year and selected the 26 highest values across the sample. A single year might have more than one of the 26 heaviest. Each of those can be thought of as 1-in-5yr events. If there are no trends in precipitation, then the total number of these events across all stations should be evenly spread over the years. In Fig. 6.4.2 we show the distribution in time of these events for the Pacific coast. The deluges associated with the massive 1997-98 El Niño event are readily apparent. While erratic, as is typical of such precipitation metrics, there is no indication of a tendency to become more frequent over time.

62

Figure 6.4.2. The time distribution by 5-year periods of the 26 heaviest (1-in-5 yr) occurrences for 29 stations on the Pacific coast.

Figure 6.4.3. As in Fig. 6.4.2 but for the heaviest 30 (1-in-5yr) events for 24 stations in the humid Southeast from Austin TX to Washington DC in 5-year bins for 1875-2024.

63

Southeast heavy rainfall events •

The trend in average precipitation is positive and statistically significant in Mobile AL and Quitman GA but insignificant elsewhere.

•

The trend in rainfall variance is positive and significant in Mobile AL but insignificant elsewhere.

•

The trend in daily maximum precipitation is positive and significant in Vicksburg MS and Norfolk VA but insignificant elsewhere.

•

Averaged over all stations in the region none of these three trend parameters is statistically significant

Figure 6.4.3 is analogous to Figure 6.4.2 for the last 150 years in the Southeast humid zone. The temporal pattern of 5-day totals of the 1-in-5yr heavy events is generally unremarkable, though a cluster of higher values appears in 1995 to 2019. The increase in those years is due largely to the 4 northeastern-most stations of Wilmington NC, Weldon NC, Washington DC, and Norfolk VA. This confirms the pattern indicated in NCA4 and NCA5 -- an increasing frequency of heavy events due to a temporal clustering of tropical storms from eastern NC to Maine discussed below. Otherwise, the remaining 20 stations show an unremarkable temporal distribution of heavy events. Northeast heavy rainfall events •

The trend in average precipitation is positive and statistically significant in 12 of 27 locations and also in the regional average.

•

The trend in rainfall variance is positive and significant in Portland ME, Albany NY, Buffalo NY and Eastport ME but insignificant elsewhere.

•

The trend in daily maximum precipitation is positive and significant in Portland ME, Gardiner ME and Eastport ME but insignificant elsewhere.

•

When averaged over all stations in the region, there is no statistically significant trend in either the precipitation variance or maximum

Fig. 6.4.4 is analogous to Figure 6.4.2 for the last 135 years in 27 stations in the Northeast (including Montreal Canada). We use 3-day totals here as this produced the largest temporal variations in time. In this region, 77 percent of events occur during June to October and are dominated by incursions of hurricanes, tropical storms, or tropical storms that transition to extratropical systems. According to NCA4 and NCA5 this region experienced the largest increases in extreme events, so it merits a closer examination. There is a noticeable clustering of extreme events from 1995 to 2014. Howarth et al. (2019) examined a similar region as in Fig. 6.4.4 that includes PA and NJ and reported significant differences in various precipitation extremes between two 18-year periods, 1979-96 and 1997-2014. That included a 317 percent increase in 24-hr events exceeding 6 inches, while we find a 58 percent increase over the same years. However, Figure 6.4.4 shows that frequency drops sharply after 2014, returning to the long-term average in the subsequent 5-year intervals, again illustrating the perils of drawing conclusions from short-term trends in highly variable metrics.

64

Figure 6.4.4. As in Fig. 6.4.2 but for the heaviest 27 (1-in-5yr) 3-day precipitation events for 27 Northeast stations from NY to ME, including Montreal.

Figure 6.4.5. Average amount of precipitation falling in the 1-in-5yr events for the NE stations.

The high percentage increase in the Howarth et al. sample is associated with small numbers at relatively few locations: there were only 6 in the first period and 25 in the second across 58 stations. Most of the stations did not record such an event. If there is a region-wide increase in heavy events, it should be seen in the average across all stations. Figure 6.4.5 shows the average precipitation in the 1-in-5 yr events across the NE. The trend is only +0.04 inches/decade. The highest amount, 1935-39, includes the Great New England Hurricane of 1938, one of the rare major (Category 3 or higher) hurricanes to strike the region. 65

The results in Figs. 6.4.4 and 6.4.5 thus suggest that though there was a surge in the number of events at a few locations during the 1995 to 2014 interval, there was no regional pattern and the change did not persist beyond 2014. Jong et al. (2024) document the increase of tropical influences on precipitation events in the Northeast since 1959 and concluded “The autumn extreme precipitation trend over the Northeast U.S. is primarily attributed to tropical cyclone-related events since the 1990s.” The question then becomes: “Was the temporal clustering of tropical systems in 1997-2014 which affected the Northeast a response to increasing GHGs?” Jong et al. examined CMIP-6 model output which suggests that there will be fewer such systems in the 21st century but that the intensity of the rainfall events might increase. This conjecture is not seen in Fig. 6.4.5 where the amount-per-event has remained steady over the 135-year period. There is some evidence to indicate the heaviest rainfall events might be redistributed due to the impact of urban infrastructure on the local weather (e.g., Pielke Sr. et al. 2011, Zhang et al. 2018, Yang et al. 2024). Yang et al.

state “Cities that experience compact development tend to witness larger increases in extreme rainfall frequency over downtown than their rural surroundings, while the anomalies in extreme rainfall frequency diminish for cities with dispersed development.” While this is an important insight to consider, the effect on the specific stations used in this analysis is unknown, or at least not detectable in Fig. 6.4.5. In summary, some U.S. regions show short-duration increases in extreme precipitation events, consistent with natural variability. But analysis of long term, nationwide historical records that considers the autocorrelation properties of precipitation data does not support the claim that extreme short-duration rainfall events are becoming more frequent or intense.

6.5

Tornadoes

AR6 assesses tornado trends in the U.S. as follows: [O]bservational trends in tornadoes, hail, and lightning associated with severe convective storms are not robustly detected due to insufficient coverage of the long-term observations. There is medium confidence that the mean annual number of tornadoes in the USA has remained relatively constant. (Chapter 11, section 11.7.3, p. 1594) The monitoring of weak tornadoes has changed over time. The growth of rural populations and the increasing ability to take video with hand-held devices has led to more frequent reports of weak tornadoes that produce minimal damage. In contrast, strong to violent tornadoes have been observed more consistently over time. Note that tornado strength is measured by the damage it produces, not by the visual appearance of the funnel. Limited real-time observational capabilities in earlier decades did not prevent identification because strong to violent tornadoes leave much more damage which will be assessed later even if the tornado itself was not observed. Since statistics began in 1950, there has been a substantial decrease (by about 50%) in the number of strong to violent tornadoes as shown in Fig. 6.5.1a. To summarize, there is a noticeable downward trend in the number of severe tornadoes in the U.S. since 1950. After 1990 the number of weak tornadoes in the U.S.

has remained roughly constant; data before that are incomplete due to limited monitoring.

66

Figure 6.5.1. Annual U.S. tornado counts for (a) strong to violent tornadoes (EF3 to EF5), and (b) weak tornadoes (EF0 to EF2). Based upon NOAA Storms Prediction Center data, available at https://www.spc.noaa.gov/wcm/data/1950-2024_actual_tornadoes.csv

6.6

Flooding

Changes in floods were assessed as follows: AR6: The SREX assessed low confidence for observed changes in the magnitude or frequency of floods at the global scale. This assessment was confirmed by the AR5 report. The SR15 found increases in flood frequency and extreme streamflow in some regions, but decreases in other regions. . . [H]ydrological literature on observed flood changes is heterogeneous, focusing at regional and sub-regional basin scales, making it difficult to synthesise at the global and sometimes regional scales. (Chapter 11.5) AR6: [T]he seasonality of floods has changed in cold regions where snowmelt dominates the flow regime in response to warming (high confidence). Confidence about peak flow trends over past decades on the global scale is low. (Chapter 11.5) NCA4: Trends in extreme high values of streamflow are mixed across the United States. Analysis of 200 U.S. stream gauges indicates areas of both increasing and decreasing flooding magnitude but does not provide robust evidence that these trends are attributable to human influences (pp. 240-241)

67

The absence of detectable US-wide trends in flooding is consistent with the findings in Section 6.4 of absence of coherent changes in extreme precipitation.

6.7

Droughts Assessments of drought trends were as follows. AR6: Few AR6 regions show observed increases in meteorological drought (Section 11.9, p. 1575), AR6: Increasing trends in agricultural and ecological droughts have been observed on all continents (medium confidence), but decreases only in one AR6 region (medium confidence). Increasing trends in hydrological droughts have been observed in a few AR6 regions. (Chapter 11 Summary) NCA4: As a consequence of this increased precipitation, drought statistics over the entire CONUS have declined. (p. 233) NCA4: Recent droughts and associated heat waves have reached record intensity in some regions of the United States; however, by geographical scale and duration, the Dust Bowl era of the 1930s remains the benchmark drought and extreme heat event in the historical record (very high confidence). (p.231) SREX: From a paleoclimate perspective, recent droughts are not unprecedented, with severe ‘megadroughts’ reported in the paleoclimatic record for Europe, North America, and Australia. (p. 170)

Figure 6.7.1: Monthly percent of US classified as “Very Dry” 1895—2025. Data source: NOAA https://www.ncei.noaa.gov/access/monitoring/uspa/wet-dry/0 least squares trendline added.

As shown in Figure 6.7.1, U.S. long-term data shows an insignificantly declining trend in extreme dryness (-0.001 percent per year) 68

Kogan et al. (2020) examines a 38-year high-resolution satellite-based drought measure and concludes that global drought has not intensified and is not connected to climate change: “it is possible to state firmly that global and main grain countries’ drought area and intensity trends have not been following global climate warming since 1980’s.” In summary there is no evidence of increasing meteorological drought frequency or intensity in the U.S. or globally over recent decades.

6.8

Wildfires

The IPCC has not provided an attribution assessment of wildfires. As shown in Figure 6.8.1, global wildfire activity as measured by European Space Agency shows a downward trend in the 21 st century.

Figure 6.8.1: Global wildfire area 2001-2018. Source: From Lizundia-Loiola et al. (2021) Figure 12. The different coloured lines represent data products derived from different satellites and algorithms

Global data show that wildfire coverage is constant or declining on every continent (Samborska and Ritchie, 2024). However there is evidence that the intensity of fires in some regions is worsening (Cunningham et al. 2024) and that wildfires resulted in a net loss of global forest cover over 2001-2019 (Tyukavina et al. 2022). Active fire suppression since 1900 makes it difficult to establish a natural baseline for wildfire activity in the U.S. Paleoclimatic evidence indicates that past activity was much higher than today. Marlon et al. (2012) used sedimentary charcoal layers to reconstruct fire history of western U.S. for the past 1400 years and also fit a model to predict fire activity as a function of climatic conditions. Their results are summarized in Figure 6.8.2 below (from Figure 2 in their paper). There has been a growing wildfire deficit over the 20th century. In other words, however much fire was observed in the 20 th century, it was less than what would have been observed in previous centuries based on the climatic conditions. Parks et al. (2025) likewise find that despite the recent increase in wildfire burn area in North America, a significant wildfire deficit remains relative to historical wildfire regimes.

69

Figure 6.8.2 Fire frequency and fire deficit in the US. The red line shows the smoothed charcoal record and the black dotted line shows the predicted charcoal record based climatic conditions. model. Source: Marlon et al. (2012) Figure 2.

U.S. data from the National Interagency Fire Centre (NIFC) from 1926 to 2023 are shown in Figure 6.8.3. The NIFC has removed the pre-1960 data from its current website on the grounds that measurement methods changed after 1960 making the comparison unreliable. Nonetheless just focusing on the post-1985 interval the number of fires is not increasing. The area burned did increase but only until about 2007. Forest fires have always been a part of nature, and they can certainly create conditions that are inhospitable in the short term for all life, including humans. Science has confirmed the overall benefit and necessity of forest fires. While recent high-profile fires and seasons serve as a reminder of the potential destructive impact, the highest profile U.S. forest fire remains the 1910 Big Blowup fire in the U.S. west, which destroyed over three million acres and eliminated entire towns like Taft, MT (Apple, 2020). The 1910 fire reshaped the U.S. Forest Service (National Forest Foundation 2022) leading to a focus on fire suppression with a primary goal of defeating all forest fires (Forest History Society, 2022). This led to the “10 am rule” in 1935 requiring that all fires spotted on any day had to be controlled by 10 am the following day (National Forest Foundation, 2022).

While defeating all fires seemed a noble goal, questions began to arise as to whether this behavior “followed the science” (U.S. Forest Service, 2022). Over time the U.S. Forest Service has begun to rethink its goals, recognizing that new approaches such as prescribed burns, fuel elimination, and controlled wildfires are more appropriate (Sommer, 2016). Recent research is validating this approach and recognizing that more frequent smaller fires likely result in healthier forests, water ecosystems and biodiversity (Stephens et al., 2021).

70

Figure 6.8.3: U.S. wildfires 1926 to 2023. Source: Post-2018: National InterAgency Fire Center data https://www.nifc.gov/fire-information/statistics/wildfires. Pre-2017 webarchive.org (n.d.).

References Apple, C. (2020, January 21). 1910 fire – The big burn across Montana and Idaho. The SpokesmanReview. https://www.spokesman.com/stories/2020/jan/19/1910-fire-big-burn-across-montana-andidaho/ Bell, Gerald D. and Muthuvel Chelliah, “Leading Tropical Modes Associated with Interannual and Multidecadal Fluctuations in North Atlantic Hurricane Activity,” Journal of Climate 19, no. 4 (February 15, 2006): 590–612, https://doi.org/10.1175/JCLI3659.1. Brewer, W. H. (1930). Up and down California in 1860–1864: The journal of William H. Brewer. University of California Press. Christy, J. R., Norris, W. B., and McNider, R. T. (2009). Surface temperature variations in East Africa and possible causes. Journal of Climate, 22(12), 3342–3356. https://doi.org/10.1175/2008JCLI2726.1 Cohn, T. A. and Lins, H. F. (2005). Nature’s style: Naturally trendy. Geophysical Research Letters, 32(23). https://doi.org/10.1029/2005GL024476 Colorado State University. (2025). Realtime tropical cyclone guidance. https://tropical.atmos.colostate.edu/Realtime/index.php Cunningham, C.X., G.J. Williamson, and D. Bowman (2024) Increasing frequency and intensity of the most extreme wildfires on Earth. Nature Ecology and Evolution 8, 1420–1425. https://doi.org/10.1038/s41559-024-02452-2 Forest History Society (2022). The 1910 fires.

https://foresthistory.org/research-explore/us-forest-servicehistory/policy-and-law/fire-u-s-forest-service/famous-fires/the-1910-fires/ Gershunov, A., and K. Guirguis (2012). California heat waves in the present and future. Geophysical Research Letters, 39(18). https://doi.org/10.1029/2012GL052979 Gershunov, A., T. Shulgina, F.M. Ralph, D.A. Lavers J.J. and Rutz (2017). Assessing the climate-scale variability of atmospheric rivers affecting western North America. Geophysical Research Letters, 44(15), 7900–7908. https://doi.org/10.1002/2017GL074175 Goldenberg, Stanley et al., “The Recent Increase in Atlantic Hurricane Activity: Causes and Implications,” Science 293, no. 5529 (July 20, 2001): 474–479, https://doi.org/10.1126/science.1060040. 71

Howarth, M. E., C.D. Thorncroft, and L.F. Bosart (2019). Changes in extreme precipitation in the Northeast United States: 1979–2014. Journal of Hydrometeorology, 20(4), 673–689. https://doi.org/10.1175/JHM-D-18-0155.1 Hurst, H. E. (1951). Long term storage capacities of reservoirs. Transactions of the American Society of Civil Engineers, 116, 776–808. Intergovernmental Panel on Climate Change (IPCC). (2021). Climate change 2021: The physical science basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (V. Masson-Delmotte et al., Eds.). Cambridge University Press. https://www.ipcc.ch/report/ar6/wg1/ Jong, B.-T., H. Murakami, T.L. Delworth and W.F. Cooke (2024). Contributions of tropical cyclones and atmospheric rivers to extreme precipitation trends over the Northeast US. Earth’s Future, 12, e2023EF004370. https://doi.org/10.1029/2023EF004370 Karl, T.R., H.F. Diaz, and G. Kukla. (1988): Urbanization: Its detection and effect on the United States climate record. Journal of Climate, 1, 1099-1123 https://journals.ametsoc.org/view/journals/clim/1/11/15200442_1988_001_1099_uidaei_2_0_co_2.xml Karl, T. R., C.N. Williams Jr., F.T. Quinlan, and T.A. Boden (1990). United States Historical Climatology Network (HCN) serial temperature and precipitation data (Publication No. 3404). Oak Ridge National Laboratory.

Klotzbach, Phil et al. (2018) Continental U.S. Hurricane Landfall Frequency and Associated Damage: Observations and Future Risks, Bulletin of the American Meteorological Society 99, no. 7 (July 2018): 1359–1376, https://doi.org/10.1175/BAMS-D-17-0184.1. Kogan, Feliz, Wei Guob and Wenze Yang (2020) “Near 40-year drought trend during 1981-2019 earth warming and food security” Geomatics, Natural Hazards and Risk Vol 11(1) Koutsoyiannis, D. (2013) Hydrology and Change. Hydrological Sciences Journal 58(6) http://dx.doi.org/10.1080/02626667.2013.804626 Lizundia-Loiola, J., et al. (2020) A spatio-temporal active-fire clustering approach for global burned area mapping at 250 m from MODIS data. Remote Sensing of the Environment 236, 111493, https://doi.org/10.1016/j.rse.2019.111493 Markonis, Y., and D. Koutsoyiannis (2016). Scale-dependence of persistence in precipitation records. Nature Climate Change, 6, 399–401. https://doi.org/10.1038/nclimate2894 Marlon, Jennifer, Patrick J Bartlein, Daniel G. Gavin et al. (2012) Long-term perspective on wildfires in the western USA Proceedings of the National Academy of Sciences February 14 2012 https://www.pnas.org/doi/pdf/10.1073/pnas.1112839109 Maue, R. N. (2025). Climatlas: Tropical cyclones. https://climatlas.com/tropical/ Maue, R. N. (2011) Recent historically low tropical cyclone activity.