by Judith Curry
Part II: what causes variations and changes in hurricane activity?
4. Detection and attribution of changes in hurricane activity
If oceans are getting warmer as a result of climate change, so the argument goes, surely hurricane activity must increase as a result, particularly hurricane intensity. However, most of the assessment reports cited in Chapter 1 have low confidence in attributing any recent changes in hurricane activity to manmade global warming.
What is the scientific basis for assessing whether or not manmade global warming is causing a change in hurricane activity?
Detection and attribution of manmade signals in in the climate system is a new and rapidly developing field. Attributing an observed change or an event partly to a causal factor (such as manmade climate forcing) normally requires that the change first be detected. A detected change is one that is determined, based on observations to be very unlikely to occur (less than about a 10% chance) due to natural internal variability alone. An attributable change implies that the relative contribution of causal factors has been evaluated, along with an assignment of statistical confidence.
There are some situations whereby attribution without detection statements can be appropriate, although lower confidence is assigned when attribution is not supported by a detected change. For example, a trend analysis for an extremely rare event may not be meaningful. Including attribution without detection in the analysis of climate change impacts reduces the chance of a false negative -incorrectly concluding that climate change had no influence on a given extreme events. However, attribution without detection comes at the risk of increasing the rate of false positives – incorrectly concluding that manmade climate change had an influence when in fact it did not.
The conceptual framework for most detection and attribution analyses consists of four elements:
- time history of relevant observations
- the estimated time history of relevant climate forcings (such as greenhouse gas concentrations or volcanic activity)
- an estimate of the impact of the climate forcings on the climate variables of interest
- an estimate of the internal (unforced) variability of the climate variables of interest—e.g. natural unforced variations of the ocean, atmosphere, land, cryosphere, in the absence of external forcings.
Paleoclimate proxies from the geological record are useful for detection studies in providing a baseline against which to compare recent variability of the past century or so. Time of emergence is the time scale on which climate change signals will become detectable in various regions – an important issue, since natural variability can obscure forced climate signals for decades, particularly for smaller space scales.
4.1 Detection
There are three main challenges to detecting a signal of changed hurricane activity:
- very long timescales in the oceans, resulting in substantial lag time between external forcing and the realization of climate change and its impacts
- high-amplitude natural internal variability in the ocean basins on time scales from the interannual to the millennial
- strong regional variations, both in ocean circulation patterns and hurricane activity.
Based on the observations summarized in Chapter 3, the following summary is provided regarding the detection of changes in global or regional hurricane activity:
- global hurricane activity: small but insignificant trends of decreasing hurricane frequency and increasing number of major hurricanes;
- global % of Category 4/5 hurricanes: increasing trend since 1970, although the data quality for the period before 1988 is disputed.
- rate of intensification: hints of a global increase, although data sets disagree.
- track migration: poleward migration of the average latitude where hurricanes have achieved their lifetime-maximum intensity for 1982-2012.
- Atlantic hurricanes: increasing trends since 1970, but comparable activity was observed in the 1950’s-1960’s.
- hurricanes in other regions: observational record is too short, but no evidence of trends that exceed natural variability
The observational database (since 1970 or even 1850) is too short to assess the full impact of natural internal variability associated with large-scale ocean circulations. Paleotempestology analyses indicate that recent hurricane activity is not unusual.
4.2 Sources of variability and change
In the absence of detecting any significant change in global or regional hurricane activity, attribution methods without detection statements must be used. These methods require assessing the contributions to climate variability/change from external forcing (e.g. CO2, volcanoes, solar) plus natural internal variability associated with the large-scale ocean circulations. In two-step attribution methods, changes in intermediate variables (such as sea surface temperature, wind shear, atmospheric humidity) are useful in identifying physical mechanisms whereby warming might contribute to a change in hurricane activity.
The focus in this section is on identifying sources of variability and change during the period since 1850, when historical data is available.
Many of the arguments surrounding an increase in hurricane activity are associated with increases in global sea surface temperature. Figure 4.1 shows the variability of globally-averaged sea surface temperature (SST) since 1850, along with external forcing from CO2, volcanoes, and the sun.
It is seen from Figure 4.1a that sea surface temperature (SST) reached a global low point in 1910, and then increased rapidly until about 1945. The elevated Atlantic hurricane activity in the 1930’s-1950’s (Section 3.3) occurred when the global SSTs were ~0.8oC cooler than present global average SSTs. This warming period was followed by a period of slight cooling until 1976, after which temperatures began increasing.
Figure 4.1 (top) Ocean surface temperature anomalies (oC) From HadSST. (middle) Human-caused carbon dioxide emissions Source: IPCC AR5. (bottom) Annual mean time series of climate forcing agents: atmospheric CO2 concentration, stratospheric aerosols (volcanic eruptions), total solar irradiance, and tropospheric aerosols Source: Hegerl et al. (2018)
The global ocean warming during the period 35-year period from 1910 to 1945 of 0.6oC was comparable to the 0.7oC warming observed between the 42-year period between 1976 and 2018.
Regarding the recent warming, the IPCC AR5 made the following attribution statement:
- “It is extremely likely that more than half of the observed increase in global average surface temperature from 1951 to 2010 was caused by the anthropogenic increase in greenhouse gas concentrations and other anthropogenic forcings together. The best estimate of the human- induced contribution to warming is similar to the observed warming over this period.”
In other words, the IPCC AR5 best estimate is that all of the warming since 1951 has been caused by humans.
So, what caused the early 20th century global warming? This issue has received remarkably little attention from climate scientists. Lack of an explanation for the early 20th century global warming diminishes the credibility of the attribution statement for warming since 1951.
The first substantive attribution analysis of the early 20th century warming was made in a recent paper by Hegerl et al. (2018), which came to the following conclusion:
“Attribution studies estimate that about a half (40–54%) of the global warming from 1901 to 1950 was forced by a combination of increasing greenhouse gases and natural forcing, offset to some extent by aerosols. Natural variability also made a large contribution. The exact contribution of each factor to large-scale warming remains uncertain.”
Hegerl et al. (2018) provides a summary of forcing from CO2, volcanoes and solar (Figure 4.1c). In 1910, the atmospheric CO2 concentration has been estimated to be 300.1 ppm; in 1950 it was 311.3 ppm; and in 2018 it is 408 ppm. So, the warming during the period 1910-1945 was associated with a CO2 increase of 10 ppm, whereas a comparable amount of warming during the period 1950 to 2018 was associated with a 97 ppm increase in atmospheric CO2 concentration almost an order of magnitude greater CO2 increase for a comparable amount of global ocean warming.
Clearly, there were other factors in play besides CO2 emissions in the early 20th century global warming (Figure 4.1b). In terms of external radiative forcing, a period of relatively low volcanic activity during the period 1920-1960 would have a relative warming effect, although the period from 1945 to 1960 was a period of slight overall cooling. Solar forcing in the early 20th century is uncertain, with estimates of warming of varying magnitude, although the magnitudes are insufficient for solar to have been a major direct contributor to the early 20th century global warming.
Hegerl et al. (2018) analyzed the internal variability associated with ocean circulations during the period since 1900. They found that the unusual cold anomaly circa 1910 (Figure 4.1a) originated in the South Atlantic, and then spread globally in the subsequent decade, leading to cold anomalies in both Atlantic and Pacific.
This rarely-discussed cold period was followed by strong warming in the Northern Hemisphere, which was particularly pronounced in high latitudes. Hegerl et al. summarized some previous research that might account for mechanisms of the strong high latitude warming in the Northern Hemisphere, including multi-decadal ocean oscillations in large-scale ocean circulation patterns. However, gaps in data coverage particularly in the Indo-Pacific and Southern Oceans imply higher uncertainty during the early 20th century than for recent periods.
Hegerl et al. focus their arguments regarding internal variability associated with large-scale ocean circulations on the Atlantic Multidecadal Oscillation and the Pacific Decadal Oscillation. The Atlantic Multi-decadal Oscillation (AMO) is a coherent mode of natural variability of sea surface temperatures (SST) occurring in the North Atlantic Ocean, with an estimated period of 60-80 years. The Pacific Decadal Oscillation (PDO) is a recurring pattern of ocean-atmosphere climate variability of surface temperature centered over the northern hemisphere mid-latitude Pacific basin. Warm phases of the both the AMO and PDO contributed to warming particularly during the 1930’s and 1940’s.
As summarized in NCA4 (2017), observed multidecadal variability in the Atlantic Ocean surface temperatures has also been ascribed to Saharan dust outbreaks and manmade pollution aerosols.
4.3 Natural internal modes of variability
Disentangling the complex interplay between the many modes of internal variability associated with the large-scale ocean circulations is not at all straightforward. The multi-decadal modes (with timescales of 30 to 80 years) are of the greatest relevance in attribution analyses of 20th and early 21st century climate change. These multi-decadal modes are associated with regional-to-basin-scale oceanic circulation systems that define the dynamical memory of the climate system in the presence of fast, large-scale atmospheric processes. The faster atmospheric processes not only supply energy for the multi-decadal variability, but also provide the means for communication between the different ocean basis and synchronization of the multi-decadal climate modes (e.g. Wyatt and Curry, 2013; Kravtsov et al. 2018).
Hurricane activity is also influenced by multidecadal variability in the oceans in ways that do not directly rely on local changes in sea surface temperatures – such as changes in atmospheric circulation patterns and wind shear.
4.3.1 Atlantic modes and hurricane activity
Three modes of interannual to multi-decadal variability have been identified in the Atlantic:
- Atlantic Multidecadal Oscillation (AMO)
- North Atlantic Oscillation (NAO)
- Atlantic Meridional Mode (AMM)
Grossman and Klotzbach (2009) provide the following summary of the relationships among these three Atlantic modes. The cross-equatorial pattern associated with the AMM and the SST, sea level pressure (SLP) and wind patterns associated with the AMO can be viewed as one overall phenomenon that stretches from the high latitudes to the tropics. The AMO and AMM are also closely related to the NAO on multidecadal time scales. Long-term positive (negative) phases of the NAO coincide with the negative (positive) phase of the AMO and AMM, generally with a lag of several years. The NAO depends on the North Atlantic meridional temperature and pressure gradient, which in turn lessens (increases) as the North Atlantic warms (cools) with the positive (negative) AMO.
The most thoroughly studied of these modes with respect to Atlantic hurricanes is the AMO. The Atlantic Multidecadal Oscillation (AMO) is associated with basin-wide SST and sea level pressure (SLP) fluctuations. The positive (warm) AMO phase is associated with a pattern of horseshoe-shaped SST anomalies in the North Atlantic Figure 4.3), with pronounced warming in the tropical and parts of the eastern subtropical North Atlantic, an anomalously cool area off the U.S. East Coast, and warm anomalies surrounding the southern tip of Greenland.
Figure 4.3 Horseshoe pattern of the AMO, where the ‘Arc’ Index corresponds to the average sea surface temperatures inside the black contours. Source: Johnstone (2017).
The traditional AMO index (Figure 4.4) is calculated from the patterns of SST variability in the North Atlantic once a linear trend has been removed. However, since the trend is significantly non-linear in time (Figure 4.1a), the detrending aliases the AMO index. The nonlinearity is particularly pronounced during the period 1945-1975, when global sea surface temperatures showed a slight cooling trend.
Figure 4.4. The Atlantic Multidecadal Oscillation (AMO) index showing positive (red) and negative (blue) phases. Source: https://commons.wikimedia.org/wiki/File:Atlantic_Multidecadal_Oscillation.svg
To avoid the problems associated with detrending, Johnstone (2017) developed an Arc Index version of the AMO Index, which is the average SST in the Arc region (Figure 4.3). The Arc Index (Figure 4.5) shows abrupt shifts to the warm phase in 1926 and 1995, consistent with the conventional AMO analysis in Figure 4.5. Johnstone’s analysis indicates a shift to the cold phase in 1971, which differs from the analysis shown in Figure 4.5 that indicates the shift to the cold phase in 1964. The revised AMO index of Klotzbach and Gray (2008) indicates a shift to the cold phase in 1970, consistent with the analysis of Johnstone.
Figure 4.5 Arc Index version of the AMO. Source: updated from Johnstone (2017)
The main hurricane-relevant variables that change with the phase changes of the AMO, AMM and NAO are spatial patterns of SST (or oceanic heat content) and wind patterns. Hurricane genesis (formation) locations, tracks and intensification are temporally and spatially modulated by these large‐scale climate modes.
Atlantic hurricanes show strong variations on decadal and multidecadal time scales in the observed record (Figures 3.6 – 3.8). The greatest impact of the AMO is on the number of major hurricanes (Category 3+) and Accumulated Cyclone Energy, shown in Figure 4.6. The shift to the relatively inactive phase occurred around 1970/1971, in accord with the AMO analyses of Johnstone (2017) and Klotzbach and Gray (2008), with the late 1960’s still characterized by a larger number of major hurricanes and high ACE values. The relationship of the AMO to major hurricane activity in the Atlantic was identified by Goldenberg et al. (2001) to be associated with above normal SSTs and decreased vertical shear associated with the warm AMO.
Bell and Chelliah (2006) related the interannual and multidecadal variability of hurricane activity in the Atlantic to two tropical multidecadal modes in the Atlantic. Comparing periods of high activity in the Atlantic, they showed that the most recent increase in hurricane activity is related to the exceptionally warm SSTs in the Atlantic, while the high activity period in the 1950s and 1960s was more closely associated with the West African monsoon.
Figure 4.6 Observations of Atlantic hurricane activity since 1920. The warm phase of the Atlantic Multidecadal Oscillation is indicated by orange shading, with the cool phase indicated by purple shading. Top: Annual frequency of major hurricanes. Bottom: Annual frequency of Accumulated Cyclone Energy (ACE). Curry (2018c)
Lin et al. (2019) argue that there are two regimes of the AMO, which appear to be consistent with the analysis of Bell and Chelliah. Lin et al. argue that there are two separate AMO regimes: a 10-30 year regime (intrinsic to the Atlantic), and a 50-80 year regime (which is influenced by variability in the Pacific and the Greenland-Iceland-Norwegian Seas).
Vimont and Kossin (2007) related Atlantic hurricane activity to the Atlantic Meridional Mode (AMM). Hurricane genesis locations, SST and wind shear anomalies are influenced by the different phases of the AMM. During the positive AMM phase (above normal SSTs in the North Atlantic), there is an overall increase of hurricane activity in the Atlantic, with the mean genesis (formation) location shifting eastward and toward the equator. Also associated with a positive AMM is an increase in storm duration and the frequency of intense hurricanes (Kossin and Vimont 2007).
4.3.2 Pacific modes and hurricane activity
The El Niño – Southern Oscillation (ENSO) is a major mode of natural climate variability. ENSO is associated with sea surface temperature (SST) changes in the tropical Pacific, which is associated with shifts in the seasonal temperature, circulation, and precipitation patterns in many parts of the world. El Niño and La Niña (warm and cold) events usually recur every 3 to 7 years and tend to last for approximately a year.
ENSO has a strong impact on hurricanes, both in the Pacific and Atlantic Oceans.
Figure 4.7 The various Niño regions where sea surface temperatures are monitored to determine the current ENSO phase (warm or cold) Source: Wikipedia
Kim et al. (2009, 2011) provide an overview of the impact of ENSO on tropical cyclones. In La Niña years, there are usually twice as many major hurricanes as in El Niño years. ENSO is generally thought to influence Atlantic hurricane activity by altering the large scale atmospheric circulation patterns for genesis (formation) and intensification. During an El Niño year, the vertical wind shear is larger than normal in most of the tropical Atlantic and especially in the Caribbean, which inhibits the formation of hurricanes.
The effect of ENSO on Pacific hurricanes is opposite to that in the Atlantic – El Nino years are associated with greater hurricane activity in the Pacific. As summarized by Kim et al. (2009a), ENSO has an impact on the mean hurricane genesis location in the Pacific, with a displacement to the southeast (northwest) in El Niño (La Niña) years. Because of this shift to the southeast, further away from the Asian continent, hurricanes in El Niño years tend to last longer and be more intense than in other years. ENSO also affects the shapes of the tracks in El Niño years, the hurricanes have a tendency to recurve northeastward and reach more northerly latitudes. Hence, hurricanes affect the southern South China Sea more frequently during La Niña years, but affect the Central Pacific more frequently in El Niño years.
Capotondi et al. (2015) address the issue of ENSO diversity, including the El Niño Modoki (a Japanese word that means ‘similar but different’). By contrast to the traditional El Niño that is associated with warming in the eastern tropical Pacific (Niño 1,2,3 regions in Figure 4.7), the El Niño Modoki is associated with warming in the central tropical Pacific (Niño 4 region). Kim et al. (2011) found that the El Niño Modoki shifts hurricane activity to the western Pacific, providing more favorable conditions for Asian landfalls, while hurricane activity in the eastern Pacific is substantially reduced. In the Atlantic, the impacts of an El Niño Modoki on hurricane activity more closely resemble a La Nina season, with elevated hurricane activity (Figure 4.8).
Figure 4.8 Composites of Atlantic track density anomaly (multiplied by 10) during the August to October period for (A) El Niño, (B) El Niño Modoki, and (C) La Niña. Source: Kim et al. (2009)
In climate change attribution studies, multi-decadal modes are of greater relevance than the interannual variability associated with ENSO and Modoki events. However, there is evidence of multidecadal variability in the relative frequency of El Niño, La Niña and Modoki events. In the Pacific, two decadal to multi-decadal modes have been identified:
- Pacific Decadal Oscillation (PDO) – an envelope for ENSO activity
- North Pacific Gyre Oscillation (NPGO) – an envelope for Modoki activity
The Pacific Decadal Oscillation (PDO) is a pattern of Pacific climate variability (poleward of 20oN), with a decadal time scale that can be interpreted as a decadal envelope of ENSO variability. During a warm (positive) phase, the west Pacific becomes cooler and part of the eastern ocean warms; during a cool (negative) phase, the opposite pattern occurs (Figure 4.9).
The North Pacific Gyre Oscillation (NPGO; DiLorenzo et al. 2008) reflects variations in the strength of the central and eastern branches of the subpolar and subtropical ocean circulation patterns, and is driven by the atmosphere through the North Pacific Oscillation (NPO). The NPO spatial pattern consists of a dipole structure in which sea level pressure (SLP) variations in the central Pacific near 40°N oppose those over Alaska. Variations of the NPGO index are shown in Figure 4.10.
Figure 4.9 PDO Index values. Source: http://research.jisao.washington.edu/pdo/
Figure 4.10 NPGO Index values. Source: https://asl.umbc.edu/hepplewhite/cindex/
DiLorenzo et al. (2015) presents an integrated hypothesis for Pacific climate variability, whereby multi-decadal patterns of Pacific decadal variability (PDO, NPGO) are energized by the interaction of meridional (north-south) modes (e.g., NPO) and the zonal (east-west) modes (e.g., ENSO and Modoki).
Maue (2011) interpreted the global Accumulated Cyclone Energy (Figure 3.2) in terms of the PDO and NPGO. The Pacific climate shifts of 1976–77 and 1988–89 have been related to the PDO and North Pacific Gyre Oscillation (NPGO), respectively, which are seen in the global ACE time series. Decadal variations in the NPGO, which has been enhanced since 1989, have been linked to SST anomaly patterns that closely resemble El Niño Modoki events.
Camargo et al. (2010) summarized several studies that have examined the decadal and multidecadal variability of hurricane activity in the western North Pacific. The observational record in the western north Pacific is unreliable before the 1950s, and perhaps even before the 1970s. The occurrence of major hurricanes is modulated by ENSO and the Pacific Decadal Oscillation. The decadal variability of hurricane tracks has also been largely attributed to the Pacific Decadal Oscillation. The regions with the greatest decadal changes are the East China Sea and the Philippine Sea.
4.3.3 Does global warming change the internal modes of variability?
The internal modes of variability associated with the large-scale ocean circulations are often referred to as ‘oscillations.’ However, it is incorrect to view these oscillations as ‘cyclic,’ as their period and frequency tend to be somewhat irregular. In principle, because they are internal modes associated with the nonlinear dynamics of the coupled atmosphere-ocean system, a specific oscillation pattern can cease to exist or change its mode of variability.
Because the historical record is relatively short, particularly outside of the Atlantic Ocean, it is useful to consider paleoclimatic evidence of these oscillations.
Knudsen et al. (2011) showed that distinct, 55- to 70-year oscillations have characterized the North Atlantic ocean-atmosphere variability over the past 8,000 years, consistent with the AMO. Cobb et al. (2013) analyzed fossil coral reconstructions of ENSO spanning the past 7000 years. The corals document highly variable ENSO activity, with no evidence for a systematic trend in ENSO variance. Twentieth-century ENSO variance is significantly higher than average fossil coral ENSO variance, but is not unprecedented. Liu et al. (2017) found that, over the period 1190 – 2007 AD, equatorial temperatures in the Central Pacific (associated with El Nino Modoki) in the late 20th century were accompanied by higher levels of interannual variability than observed previously in this period.
The NCA4 (2017; Chapter 5) concluded that confidence is low regarding the impact of manmade global warming on changes to these internal modes associated with large-scale ocean circulation patterns.
4.4 Attribution – models
Extended integrations of global climate models in principle should allow for an assessment of the frequency, intensity, duration and tracks of hurricanelike features in the model simulations. Attribution of the impacts of manmade global warming on hurricane characteristics can then be assessed through comparing climate model simulations both with and without human impacts (e.g. CO2 and aerosol emissions).
A prerequisite for using global climate models for attribution analyses or 21st century projections of hurricane activity requires that historical climate model simulations accurately simulate hurricane characteristics and interannual to decadal variability. However, simulation of realistic hurricane characteristics is hampered by the coarse resolution generally required of such global models and also the model treatment of tropical convection and clouds (e.g. Camargo et al 2008; Walsh et al. 2015). Further, climate models do not accurately simulate the timing and patterns of the multi-decadal oscillations (e.g. Kravtsov et al. 2018).
A number of new, high-resolution simulations of the generation of hurricanes by global climate models have been performed in recent years (see Walsh et al. 2016 for a summary). More realistic maximum hurricane intensities have been simulated by downscaling individual storm cases from a coarse grid global model into a regional high resolution hurricane prediction system.
As a recent example, Patricola and Wehner (2018) used a high-resolution model to simulate 15 hurricane events from the global historical record. Simulations for each storm were conducted under current climate conditions versus the surface climate associated with pre-industrial conditions. They found that, relative to pre-industrial conditions, climate change has enhanced the average and extreme rainfall of hurricanes Katrina, Irma and Maria by 4%–9% and increased the probability of extreme rainfall rates, suggesting that climate change to date has already begun to increase tropical cyclone rainfall.
The model used by Patricola and Wehner (2018) was driven by specified sea surface temperatures, and did not include coupling to the ocean. Lack of ocean coupling in the model can lead to tropical cyclones that are more intense and frequent compared to slab–ocean and fully coupled atmosphere–ocean simulations. Tropical cyclone winds typically induce a ‘cold wake’ of upper-ocean temperatures. The cold wake can reduce the tropical cyclone intensity, depending on the tropical cyclone’s intensity and translation speed and the ocean heat content and salinity structure. Further, these simulations of individual storms only include the thermodynamic (temperature) related aspects of climate change, and do not include the impact of any atmospheric or ocean circulation changes that might be associated with global warming.
In one of the most sophisticated model-based attribution studies to date, Bhatia et al. (2019) investigated the issue of whether hurricane rates of intensification are increased by global warming. They compared the observed trends to natural variability in bias-corrected, high-resolution, global coupled model experiments that accurately simulate the climatological distribution of tropical cyclone intensification. Their results suggest a detectable increase of Atlantic intensification rates with a positive contribution from manmade forcing and reveal a need for more reliable data before detecting a robust trend at the global scale. The paper concludes that the study is limited by the ability of a climate model to accurately represent natural variability as well as the uncertainty around the trends in relatively short observational records. Further analysis with additional high-resolution climate models and a longer and more reliable observational record is required to confirm these conclusions.
In summary, global climate models are currently of limited use in hurricane attribution studies. High-resolution models used to simulate individual hurricanes are being used to perform controlled experiments that focus on specific events and the complexities of relevant physical processes. However, definitive conclusions regarding the impact of manmade warming on hurricanes cannot be determined from these simulations, given the current state of model development and technology.
4.5 Attribution – physical understanding
Our knowledge of the relationships between climate variability and hurricanes comes mainly from the analysis of historical data. Meaningful interpretation of these relationships requires understanding of the mechanisms that determine these relationships, but ultimately this understanding is limited by the same fundamental factors that limit our understanding of the mechanisms of the formation and intensification of individual hurricanes (see Emanuel 2018 for a review of current knowledge of hurricane processes).
4.5.1 Genesis
While there are some theories for hurricane genesis (formation), there is no quantitative theory that relates the probability of genesis to the large-scale environmental conditions. As summarized by Camargo et al. (2008), we have known for decades that sea surface temperature, vertical wind shear, and atmospheric humidity influence genesis, and this gives us an empirical basis for understanding how climate variations influence hurricane numbers.
As summarized by Walsh et al. (2015), the number of hurricanes appears to be related to changes in the mean vertical circulation of the atmosphere. Research indicates that thermodynamic variables (related to temperature and humidity) are generally more important than atmospheric circulations for hurricane formation in the North Atlantic. Humidity in the lower atmosphere was shown to be the most important controlling parameter for formation in the Atlantic, with sea surface temperatures and cyclonic circulations patterns, wind shear and rising motion also being important.
The problem of understanding the impact of global warming on hurricane genesis is complicated by potentially compensating influences of a warming climate on hurricanes (e.g. Patricola and Wehner, 2018). Increasing sea-surface temperature (SST) are expected to intensify tropical cyclones. However, projected increases in vertical wind shear could work to suppress tropical cyclones regionally.
As summarized by IPCC AR5 (2013; Chapter 16), hurricanes can respond to manmade forcing via different and possibly unexpected pathways. For example, increasing emissions of black carbon and other aerosols in South Asia has been linked to a reduction of SST gradients in the Northern Indian Ocean, which has in turn been linked to a weakening of the vertical wind shear in the region and an observed increase in the number of intense hurricanes in the Arabian Sea. In the North Atlantic, the reduction of pollution aerosols is linked to tropical SST increases, while in the northern Indian Ocean, increases in aerosol pollution have been linked to reduced vertical wind shear – both of these effects have been related to increased tropical cyclone activity.
4.5.2 Intensification
The causal chain for global warming to increase hurricane intensity has long been argued to occur via the increase in sea surface temperature (SST) (e.g. Curry et al. 2006). Hoyos et al. (2006) showed that the trend of increasing numbers of category 4 and 5 hurricanes for the period 1970-2004 is directly linked to the trend in sea-surface temperature; other aspects of the tropical environment, although they influence shorter-term variations in hurricane intensity, do not contribute substantially to the observed global trend.
A nominal SST threshold of 26.5oC [80 oF] has been used as a criterion for the formation of hurricanes, and a threshold of 28.5oC [82.4 oF] for intensification to a major hurricane (Category 3+). New insights into the relationship between warming and hurricane intensity are provided by Hoyos and Webster (2011). During the 20th century, tropical ocean SST has increased by about 0.8°, accompanied by a steady 70% expansion of the ocean warm pool area that encompasses the regions exceeding 28oC [82.4 oF]. However, the region of tropical cyclogenesis has not expanded, owing to the area of convective activity remaining nearly constant. Hoyos and Webster argue that the temperature threshold for tropical cyclogenesis increases as the average tropical ocean temperature increases. The increasing intensity of atmospheric convection with warmer temperatures seems to be the link between SST increase and hurricane intensity, rather than the absolute value of the SST itself. Further, the location of the intense convection is related to the difference between the local SST and global tropical average SST, rather than to the absolute value of the SST itself (Vecchi et al. 2008). This variation in the threshold temperature for hurricane formation and intensification is consistent with the existence of very intense hurricanes even when the climate was significantly cooler.
The causal link between SST and hurricane plays a prominent role in theories to estimate the upper bounds on tropical cyclone intensity that indicate that there is a strong relationship between ocean thermal energy and the maximum potential intensity that can be achieved. Potential intensity is defined as the maximum sustainable intensity of a hurricane based on the thermodynamic state of the atmosphere and sea surface. The theories of potential intensity continue to be challenged and refined. Knutson and Tuleya (2004) estimated the rough order of magnitude of the sensitivity of hurricane maximum intensity to be about 4% per degree C of SST warming. Such sensitivity estimates have considerable uncertainty, as shown by a subsequent assessments.
4.5.3 Rainfall
As the ocean surface warm, more water evaporates and a warmer atmosphere has a greater capacity to hold water vapor. Simple thermodynamic calculations show that there is about 7% more water vapor in saturated air for every 1°C [2 oF] of ocean warming (e.g. Trenberth, 2007).
This increase in atmospheric water vapor can cause an even larger increase in hurricane rainfall, since water vapor retains the extra heat energy required to evaporate the water, and when the water vapor condenses into rain, this latent heat is released.
4.6 Conclusions
Models and theory suggest that hurricane intensity and rainfall should increase in a warming climate. There is no theory that predicts a change in the number of hurricanes or a change in hurricane tracks with warmer temperatures.
Convincing attribution of any changes requires that a change in hurricane characteristics be identified from observations, with the change exceeding natural variability.
The global percent of Category 4/5 hurricanes has been observed to be increasing, although the amount of the increase depends on period considered, with questionable observations in some regions prior to 1987. Because of the short length of the data record, attribution of any portion of this increase to manmade global warming requires careful examination of the data and modes of natural variability in each of the regions where hurricanes occur.
While theory and models indicate that hurricane rainfall should increase in a warming climate, satellite-based observational analyses of hurricane rainfall have not addressed this issue on a meaningful spatial or temporal scale.
There is some evidence for a slowing of tropical cyclone propagation speeds globally over the past half century, but these observed changes have not yet been confidently linked to manmade climate change.
While substantial increases in Atlantic hurricane activity have occurred since 1970, these increases are likely driven by changes in the Atlantic Multidecadal Oscillation (AMO) and Atlantic Meridional Mode (AMM). Climate model simulations suggest a recent increase in the rate of intensification of Atlantic hurricanes that exceeds what can be expected from natural internal variability.
If manmade global warming is causing an increase in some aspect of hurricane activity, this increase should be evident globally, and not just in a single ocean basin. One problem is that data is insufficient for detection on the global level. When considering a single ocean basin, correct interpretation and simulation of natural internal variability is of paramount importance; unfortunately our understanding and ability to correctly simulate natural internal variability with global climate models is limited.
In summary, the trend signal in hurricane activity has not yet had time to rise above the background variability of natural processes. Manmade climate change may have caused changes in hurricane activity that are not yet detectable due to the small magnitude of these changes compared to estimated natural variability, or due to observational limitations. But at this point, there is no convincing evidence that manmade global warming has caused a change in hurricane activity.
JC note: stay tuned, next two posts will be on landfalling hurricanes.
Thanks Judith very interesting
Terrific citations.
“Models and theory suggest that hurricane intensity and rainfall should increase in a warming climate.”
I find that hard to believe. Both atmospheric and oceanographic circulations are driven by the temperature difference between the equator and the poles. We are told that the main effect of global warming is to reduce the temperature difference by raising the high latitude temperatures. In that case, I would expect that a warmer planet would have fewer and less intense hurricanes. That goes for tornadoes, too.
“I find that hard to believe. Both atmospheric and oceanographic circulations are driven by the temperature difference between the equator and the poles. We are told that the main effect of global warming is to reduce the temperature difference by raising the high latitude temperatures. In that case, I would expect that a warmer planet would have fewer and less intense hurricanes. That goes for tornadoes, too.”
You are mixing up systems spawned from the PJS (Polar Jet Stream), which derive their energy from the vorticity generated by a strong PJS, and hence strong warm air advection into them, and that of Tropical storms which get their energy from sensible and latent heat released from SSTs.
Tornadoes are driven by the CAPE available in the air-mass. Warm/moist air overlain by dry cool air. There need to be vertical shear present also, to initiate rotation. –
So no, a warmer planet (warmer SST’s) will have more energy available to fuel TCs. However the PJS will become weaker making temperate depressions less likely to explosively deepen, but on the other hand lead to more Rossby wave and cut-off low/high development with consequent ‘stuck’ weather (prolonged rain/drought).
“You are mixing up systems spawned from the PJS (Polar Jet Stream), which derive their energy from the vorticity generated by a strong PJS, and hence strong warm air advection into them, and that of Tropical storms which get their energy from sensible and latent heat released from SSTs.”
There’s evidence that a large number of tropical cyclones occur from cross equatorial polar air mass intrusions. The winter hemisphere tends to “push” the ITCZ well into the summer hemisphere ( everywhere except the SE Pacific and the S Atlantic, where tropical cyclones are very rare ):
https://image.slidesharecdn.com/winds-111208234718-phpapp02/95/winds-16-728.jpg
This is probably an effect of formation more than intensity, but the existence of intense cyclones depends on the intensification of fledgling cyclones. And there doesn’t seem to be much effect on inter-hemispheric RF from global warming. But the effect of the polar circulation may well be important.
“Tornadoes are driven by the CAPE available in the air-mass. Warm/moist air overlain by dry cool air. There need to be vertical shear present also, to initiate rotation. –”
I think you will find that the overwhelming majority of tornadoes occur in association with cold fronts, usually marked by strong temperature gradients and strong polar jet streams.
“So no, a warmer planet (warmer SST’s) will have more energy available to fuel TCs. However the PJS will become weaker making temperate depressions less likely to explosively deepen, but on the other hand lead to more Rossby wave and cut-off low/high development with consequent ‘stuck’ weather (prolonged rain/drought).”
This is an interesting theory, but it is not yet supported by observation.
And there is some contrary theory. To be sure, at the surface, and above, so called Arctic Amplification tends to raise winter Arctic temperature more, weakening the surface and lower temperature gradient, reducing wintertime thermal wind speeds.
But recall that most models of AGW create a “Hot Spot” of warming.
From 500mb to 100mb, coincident with jet streams, this feature would act to greatly increase the temperature gradient, intensifying jet streams.
The hot spot remains unvalidated, so…
But, AGW related jet stream change may never be observed because jet stream undulations themselves change temperature gradients, amplifying in some areas, weakening in others. Variables which depend on themselves are inherently non-linear. Year-to-year variations in jetstream location and intensity are large and so may well render any AGW influence moot.
“This is an interesting theory, but it is not yet supported by observation.”
It is observed by the meteorological community every NH summer.
The PJS is much weaker then and cut-off lows/highs are much more frequent.
It is entirely logical that further weakening of the mid lat to Pole DeltaT and therefore of the summer PJS, would only make “stuck” Wx systems more likely.
https://www.nature.com/articles/s41467-018-05256-8
“I think you will find that the overwhelming majority of tornadoes occur in association with cold fronts, usually marked by strong temperature gradients and strong polar jet streams.”
Yes, they do, as that is where the triggers I mentioned as necessary are most likely to be found.
“The hot spot remains unvalidated, so…”
It has been found, and is a function of LH release over Tropical ocean.
No point in linking to the latest studies as you have been shown them. Multiple times and continue to naysay them.
“But, AGW related jet stream change may never be observed because jet stream undulations themselves change temperature gradients, amplifying in some areas, weakening in others. ”
Indeed – so you end up with a mean (just as now) and the Polar amplification on top which still will weaken the overall DeltaT.
“But recall that most models of AGW create a “Hot Spot” of warming.
From 500mb to 100mb, coincident with jet streams, this feature would act to greatly increase the temperature gradient, intensifying jet streams.”
No, we are talking of the ITCZ – that air is deflected by Coriolis and creates the Tropical Hadley cell as it at first diverges aloft and then converges and sinks at ~ 30 deg N/S to form the sub-tropical high pressure belts.
A JS can have 10’s of degrees of horizontal DetaT as a cross-section through its core level. The hot-spot barely a degree.
“The hot spot remains unvalidated, so…”
It has been found, and is a function of LH release over Tropical ocean.
No point in linking to the latest studies as you have been shown them. Multiple times and continue to naysay them.
As humans, we all have confirmation bias. I’m beginning to appreciate that it may even have been evolutionarily beneficial in that confidence, even if the result of erroneous assumptions, may benefit us more than the paralysis of uncertainty.
However, we can point out bias to one another, and hopefully better assess reality. As is easily and repeatably demonstrable from:
RSS MSU data
UAH MSU data
NOAA STAR MSU data
RATPAC RAOB and
IGRA complete RAOB,
the modeled hot spot is not evident for the period from 1979 to present.
Some gravitate toward the Sherwood analysis of raob data, which does least indicate increased warming with height in the hot spot regions.
The Sherwood analysis: 1.) is the outlier wrt hot spot; 2.) uses kriging to perhaps inappropriately homogenize RAOB data; 3.) includes stations with fragmentary records which are of questionable quality; and 4.) is out of date
It could still be that the Sherwood analysis is the most accurate – there is always uncertainty. However, confirmation bias is choosing data which confirm pre-existing ideas while dismissing data which contradicts our ideas. For the hot spot, one must employ a lot more confirmation bias to cling to the minority analysis.
So, why would it matter?
With respect to the topic at hand, a hot spot (greater warming above lesser warming ) would mean increased static stability of the tropical cyclone regions, which should tend to reduce Convective Available Potential Energy. This is certainly contrary to the ideas of increased tropical cyclone energy.
Tony Banton,
There is less wind when the planet is warmer. Look at the dust content in ice cores over the past 800 ka. Dust content is high when the planet is cold and less when warmer, implying more wind when cold.
Congratulations Prof. Curry for presenting a reasonably comprehensive coverage of the main decadal and inter-decadal factors that could play a role in generating hurricane/typhoon/cyclone activity.
Unfortunately, your review completely ignores a factor that I believe has a profound effect on the generation of hurricane/typhoon/cyclone activity.
It is easy to follow the generation of nascent tropical low-pressure cell pairs straddling the equator, associated with the active phase of Madden Julian Oscillations (MJOs – over the Indian Ocean and Western Pacific) and with easterly propagating Equatorial Kelvin Waves (EKWs – over the rest of the equatorial regions). I do not think that it is a stretch to conclude that the generation of these nascent low-pressure cell pairs plays a role in the subsequent development of hurricanes, typhoons, and cyclones.
The physical phenomena that generate the low-pressure cell pairs straddling the Equator are the westerly propagating Equatorial Rossby Waves (ERWs) that trail behind the easterly propagating MJOs and EKWs.
The question that people should be asking is what is the physical mechanism that spawns the period appearance of ERWs behind the MJOs and EKWs?
I know that you are a very busy person, so I understand that you almost certainly do not have the time to read the following blog post that I have made on the underlying physical mechanism,
http://astroclimateconnection.blogspot.com/2019/02/the-lunar-tidal-model-part-4.html
however, you may know of someone who could read it and give you an executive summary.
Click on the link below if you want to see a preliminary test of the following hypothesis.
https://astroclimateconnection.blogspot.com/2019/02/the-north-atlantic-hurricane-season.html
Hypothesis
Tropical depressions or storms that appear in the Atlantic Ocean between the Equator and 23.5 degrees North during the 2016 North Atlantic Hurricane season, will do so on dates that are maxima or minima in the lunar-induced changes in the relative angular velocity of the Earth’s rotation. [N.B. the dates that are maxima or minima in the lunar-induced changes in the relative angular velocity occur close to the times when the Moon crosses the Earth’s equator or reaches lunar standstill (i.e. the Moon is furthest north or south of the Equator).]
This sample used to test this hypothesis includes all tropical depressions/storms that start in the Atlantic Ocean between latitudes of 0 degrees and 23.5 degrees North. [Note: tropical depression 9 – Hurricane Herman is included as it is a borderline case since it is first generated at 23.8 N]. The reason for excluding tropical depressions and tropical storms that start further than 23.5 degrees away from the equator is that their generation may have more to do with factors related to the mid-latitudes rather than the equatorial regions.
All but 2 of the 12 tropical depressions and storms that occurred during the North Atlantic 2016 Hurricane season support the proposed hypothesis.
For a more detailed discussion of the reasons for proposing this hypothesis please read:
https://astroclimateconnection.blogspot.com/2019/02/the-lunar-tidal-model-part-4.html
Here are the results for the North Atlantic tropical storms Colin and Danielle which occurred in June of 2016.
KEY FOR FIGURE
TD = Tropical Depression
TS = Tropical Storm
CATN = Category N Hurricane where N = 1, 2, 3, 4, or 5.
PK = Peak Activity
https://2.bp.blogspot.com/-82dIrWvUq9U/XG5895a6ekI/AAAAAAAABiw/NuKpo-Iky3gOujftz24CvNzj8CduFZ5pACLcBGAs/s640/AV_JUNE_2016.jpg
It looks as though all but one (the exception being Katia) of the 15 Topical Depressions/Storms/Hurricanes that occurred during the 2017 North Atlantic Hurricane season supports the claim that:
Tropical depressions or storms that appear in the Atlantic Ocean between the Equator and 25.0 degrees North during the North Atlantic Hurricane season, will do so on dates that are maxima or minima in the lunar-induced changes in the relative angular velocity of the Earth’s rotation. [N.B. the dates that are maxima or minima in the lunar-induced changes in the relative angular velocity occur close to the times when the Moon crosses the Earth’s equator or reaches lunar standstill (i.e. the Moon is furthest north or south of the Equator).]
https://astroclimateconnection.blogspot.com/2019/02/north-atlantic-hurricane-season-june-to.html
This claim is also supported true by 10 of the 12 Topical Depressions /Storms/Hurricanes that occurred during the Hurricane season 2016.
KEY FOR FIGURES
Pot# = Potential Tropical Cycle number #
TD = Tropical Depression
TS = Tropical Storm
CATN = Category N Hurricane where N = 1, 2, 3, 4, or 5.
PK = Peak Activity
Regen. = Regenerated
https://2.bp.blogspot.com/-6loC1n5BL9s/XG-t6Xgw2TI/AAAAAAAABko/T8hwfcyNoVAOAO1d_wY6dbtDPslL4yKDACLcBGAs/s1600/AV_SEP_2017.jpg
Reblogged this on Climate Collections.
Breaking news: “White House prepares to scrutinize intelligence agencies finding that climate change threatens national security” (WashPo)
https://burnmorecoal.com/2019/02/20/white-house-preparing-climate-panel/
Here’s hoping for this proposed Executive Order calling for “a rigorous independent and adversarial scientific peer review to examine the certainties and uncertainties of climate science, as well as implications for national security.”
About time.
Hi, you might be interested in our 2017 GRL paper (Defforge & Merlis) following up Johnson & Xie 2010, assessing whether the SST at TC genesis has warmed in observations: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016GL071045
Hi Timothy, thanks much for the link to your paper
I have my doubts about this IPCC statement: “It is extremely likely that more than half of the observed increase in global average surface temperature from 1951 to 2010 was caused by the anthropogenic increase in greenhouse gas concentrations and other anthropogenic forcings together.” My concern is that it is very difficult to warm water by increasing the air temperature above it. On the other hand, the intensity of sunlight has a big impact, think how the seasonal variation of sun affects water bodies. If the IPCC rationale is that the anthropogenic increases decrease cloud cover or somehow changes sunlight intensity which in turn drives SST then I could buy into the attribution argument for hurricanes. Otherwise I think this is pure speculation.
Thanks for the extensive perspective. Any comments on the recent article you “liked”?
Li, Zhibo, Tim Li, and Ying Sun. “Relative roles of dynamic and thermodynamic processes in causing positive and negative global mean SST trends during the past 100 years.” Dynamics of Atmospheres and Oceans (2019). http://buff.ly/2Ed5IK7
They find that “water induced greenhouse effect offsets the CO2 effect during hiatus phases”. i.e. atmospheric water increase or decrease caused the warming or warming haitus.
Very interesting, Judith.
Regarding this:
“The conceptual framework for most detection and attribution analyses consists of four elements:”
The problem is that this approach relies on multiple assumptions, not the least that the forcings are properly quantified.
An alternative approach is to use dynamical systems identification. System identification can bypass the need for such assumptions.
Using this approach Philippe de Larminat found a much lower anthropogenic attribution for recent global warming:
de Larminat, P. (2016). Earth climate identification vs. anthropic global warming attribution. Annual Reviews in Control, 42, 114-125.
https://www.sciencedirect.com/science/article/pii/S1367578816300931
Open article.
No wonder the IPCC prefers obscure attribution methods where everything depends on the assumptions.
Javier: Using this approach Philippe de Larminat
Thank you for the link. I liked the paper. However, the big question that remains is what to do with the long-term temperature reconstructions, as stated clearly by the authors in the abstract:
We explain the differences first by the exclusion by IPCC of the millennial paleoclimatic data, … .
They are not the first to observe how much depends on whether there is or is not a long period oscillation driving the recent increase in temperature.
Quote: “They are not the first to observe how much depends on whether there is or is not a long period oscillation driving the recent increase in temperature.” Spot on.
Javier, in ‘Nature Unbound IX’, in fig 122, has linked the long period Eddy cycle, peaks and troughs, to well known conditions in the last 2k years. The Eddy troughs I found correlated to events discerned from my earlier delving, consecutively from 6k2bce to 2k345bce. Five dates found several years ago, with some precision, from various correlated proxies including tree-rings. Hardly coincidence.
Javier above refers to ‘System identification’ – which has particular reference to ‘control systems’. Spot on again. The dynamics of the rotating Earth are based on a too simplified dynamic, and nowhere considers that other factors also are involved. Factors that have left there mark in both obliquity measurements from historical times to evidence of archaeological nature that prove the fact. That evidence fits comfortably (actually very much in opposite) with the correlated dates of the Eddy cycle.
mrm: I liked the paper.
That is faint praise — the paper repays careful reading. I recommend it, and repeat my “thank you for the link” to javier. I think any essay on attribution in CO2-Climate relationships ought to refer to it.
Judith thank you. That is wonderful scientific analysis and summary. Easy to read, logical and impressive. Smile young lady.
.
…especially when it’s ostensibly the global warming alarmists that are granting this concession to common sense.
Judith, you might submit the conclusions to The Australian’s opinion/commentary pages. The widespread ignorance down her is appalling. [I’m drafting an op-ed piece on a climate topic after contacting Graham Lloyd, but very slowly given very low energy levels and attendant lack of focus.]
PS: after a very dry “wet season” in South East Queensland, we are hoping that a lurking Cat 1/2 cyclone will come close enough to give us a good soak.
I. “… this increase should be evident globally, and not just in a single ocean basin …”
Not necessarily, I recommend the papers Ph. Sura’s (https://www.researchgate.net/scientific-contributions/81845472_Philip_Sura) especially: A general perspective of extreme events in weather and climate, (2011.).
II. Robert Sykes: “In that case, I would expect that a warmer planet would have fewer and less intense hurricanes.”
You are right, this factor can be underestimated – undervalued in models. I will remind you, for example:
Dezileau et al., 2011. Intense storm activity during the Little Ice Age on the French Mediterranean coast (http://hal.archives-ouvertes.fr/hal-00617525/): “The apparent increase of the superstorm activity during the latter half of the Little Ice Age was probably due to the thermal gradient increase …”
Trouet et al., 2012. North Atlantic storminess and Atlantic Meridional Overturning Circulation during the last Millennium: Reconciling contradictory proxy records of NAO variability
(http://www.sciencedirect.com/science/article/pii/S092181811100155X):
“Such an increase in cyclone intensity could have resulted from the steepening of the meridional temperature gradient as the poles cooled more strongly than the Tropics from the MCA into the LIA.”
Shah-Hossein et al., Coastal boulders in Martigues, French Mediterranean: evidence for extreme storm waves during the Little Ice Age
2013, (http://www.schweizerbart.de/papers/zfg_suppl/detail/57/81545/Coastal_boulders_in_Martigues_French_Mediterranean_evidence_for_extreme_storm_waves_during_the_Little_Ice_Age ):
“Dating of the boulders shows age ranges that correspond to the Little Ice Age (LIA), thus suggesting a relationship between their deposition and the high storm frequency that characterized the LIA.”
Donnelly et al., 2006. Climate Forcing of North Atlantic Tropical Cyclone Activity over the last 6000 years (http://adsabs.harvard.edu/abs/2006AGUFM.U51C..01D):
“However, given the increase in intense tropical cyclone landfalls during the later half of the LIA, tropical SSTs as warm as present are apparently not a requisite condition for increased intense …”
III. To sum up: these sentences (from various papers) by T. Knutson are still valid:
“Both the increased warming of the upper troposphere relative to the surface and the increased vertical wind shear are detrimental factors for hurricane development and intensification, while warmer SSTs favor development and intensification. To explore which effect of these effects might “WIN OUT” …”
“The projection of more frequent intense hurricanes is statistically significant for the CMIP3 ensemble climate change, but only nominally positive, and not statistically significant, for the CMIP5 ensemble.”
“… intensity projected for the Atlantic basin showed relatively small changes in some studies, ranging even to NEGATIVE VALUES for some individual models that were analyzed …”
“… the short time period of this dataset, together with the lack of “Control run” estimates of internal climate variability of TC intensities, precludes a climate change detection at this point …”
“We have very low confidence in projected changes in individual basins.”
“First, it is possible that 21st century changes in tropical cyclones will be less potentially damaging than the scenarios outlined in the projections section.” “Global climate transient sensitivity or sea level rise could be at the low end, or even lower than, the range shown in IPCC AR4.”
………………………….
semczyszakarkadiusz: I, II, III
That is a fine post. Thank you for the links.
The understanding that a warm AMO phase is a response to low solar via negative NAM, provides the frame of reference for attribution. If rising radiative forcing increases positive NAM, that should relatively cool the AMO and reduce Atlantic hurricane intensity.
IPCC AR4 Annular Modes and Mid-Latitude Circulation Changes
https://archive.ipcc.ch/publications_and_data/ar4/wg1/en/ch10s10-3-5-6.html
SST proxies by Greenland from 300 AD:
https://media.springernature.com/m685/springer-static/image/art%3A10.1038%2Fs41598-017-13246-x/MediaObjects/41598_2017_13246_Fig2_HTML.jpg
.
❶①❶①❶①❶①❶①❶①❶①❶①❶①❶①❶
❶①❶①
❶①❶① . . . Global Warming Travel Warning . . .
❶①❶①
❶①❶①❶①❶①❶①❶①❶①❶①❶①❶①❶
.
The IPCC has issued an urgent travel warning.
Many people are foolishly travelling to countries, which have an average temperature which is more than 2 degrees Celsius warmer than their home country.
This activity is highly dangerous, and could result in the deaths of millions of people.
The IPCC suggests that people limit their travel, to countries which have an average temperature which is less than 1.5 degrees Celsius warmer than their home country.
Humans evolved in Africa, many millions of years ago. Climate scientists use the abbreviation “BT”, when they refer to this time (“BT” stands for “Before Thermometers”).
In the early days, early humans never travelled more than a few miles, over their entire lifetime. They never travelled more than a few miles, because kilometres had not yet been invented.
Humans, therefore, became adapted to a very narrow temperature range. Going outside of that narrow temperature range, could be deadly. Many early humans were eaten by lions, because they went outside of their normal temperature range.
But early humans had one advantage, that the other animals didn’t have. Because they never washed, early humans tasted horrible, and they didn’t smell very nice. So the other animals left early humans alone. And humans were able to travel all over the Earth.
====================
Scientists have proved that travel and temperatures, are more dangerous than smoking 60 cigarettes a day, for 50 years.
It is safer to stay at home, and take up smoking, than to go travelling in warmer countries.
Don’t worry. We understand that humans have an “urge” to travel. It comes from our early ancestry, when we had to find large herds of animals to eat.
Here at the IPCC, we want what is best for YOU. And we have had our top scientists work out a “safe” way of travelling.
To ensure your personal temperature safety, the IPCC has emitted the following travel regulations.
Travel will be limited to “safe” country groups. This means that travel may only take place between a country, and the other countries that are in the same temperature safety group.
For further deatails, please click the following link:
https://agree-to-disagree.com/global-warming-travel-warning
Using hurdat and monthly ERSST, I looked at some correlations for the Atlantic and NE Pacific.
Three things emerge.
1. In figures b. below, the temperature ( ~28.5C Atlantic, ~28.0C NEPAC ), at which most weaker storms peak is the same as the temperature at which major hurricanes peak. There are many interdependent variables ( latitude of path, shear, landfalls, etc. ), but this does not support the idea of strong intensity dependence on temperature.
2. In figure c. below, in the Atlantic, weaker storms as well as major hurricanes peak at SSTs about the same as the formation SSTs which doesn’t indicate great temperature dependence. There is a greater percentage of major hurricanes which peak at SSTs greater than the formation SSTs, however, this also may be a consequence of more equatorward formation latitude, which spares the storm from shear for a longer duration.
3. In the NE Pacific, most major hurricanes as well as weaker storms peak SSTs lower than the formation SSTs. This has to do with the shape of the SST distribution ( which of course, may co-vary with the shear and storm paths ). But this also contra-indicates strong temperature dependence.
https://turbulenteddies.files.wordpress.com/2017/09/atlantic_1949_2016_figure_3.png
https://turbulenteddies.files.wordpress.com/2017/09/pacific_1949_2016_figure_6.png
Judy, perhaps the paper of Sounders et al. (2017) https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2017JD026492 is helpfull. They replicate the NA-hurrican ACE since 1870s with the trade winds and the SST difference MDR-global tropics. This value has no trend. Some models suggest reduced trades in a warming climate… models! I recalculated it with ERAint: no significant trend in the Atlantic trades 1979…2018 during June-September.
A friend of mine late last century found that tropical cyclones in our region were twice as frequent in La Niña than El Niño years. La Niña (El Niño) states are more frequent in cool (warm) eastern Pacific sea surface temperature epochs.
https://watertechbyrie.files.wordpress.com/2019/02/tpi-sst.png
http://joellegergis.com/wp-content/uploads/2007/01/Henley_ClimDyn_2015.pdf
ENSO epochs are behaviors that emerge and shift chaotically over millennia. El Niño frequency is at a millennial high. They can be traced in a millennial long instrumental record.
https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/1999GL900013
In corals.
http://science.sciencemag.org/content/339/6115/67
In geomorphology – although Australian cyclone frequency is at a low ebb El Niño enhances tropical cyclone frequency in the Atlantic.
https://www.nature.com/articles/nature12882
In an Antarctic ice core.
https://watertechbyrie.files.wordpress.com/2014/06/vance2012-antartica-law-dome-ice-core-salt-content-e1540939103404.jpg
https://journals.ametsoc.org/doi/10.1175/JCLI-D-12-00003.1
Detection of change in cyclone frequency require a much broader temporal and spatial perspective than is sometimes the case in climate science. Attribution may be an order of magnitude more difficult problem.
“Detection of change in cyclone frequency require a much broader temporal and spatial perspective than is sometimes the case in climate science. Attribution may be an order of magnitude more difficult problem.”
I agree completely. Much the same point can be made about both detection and attribution of virtually any aspect of CO2/manmade climate change.
Tropical equivalently cyclones, typhoons or hurricanes require warm water found close to the equator and the Coriolis effect at latitudes higher than about 8 degrees north and south. The two come together in the main development region. Sherrington’s tracks over land below are in the dissipation phase.
https://www.youtube.com/watch?v=UKL9NIxLIIE
SST in the tropical Pacific changes with ENSO – with the areal extent of warm or cold surfaces and the ocean/cloud positive feedback. The familiar decadal pattern can be seen in Judith’s Figure 4.1. Although I am now thinking that we have been mislead by the apparent regularity of 20th century variability – and that these ENSO epochs are far more deterministically chaotic.
https://watertechbyrie.files.wordpress.com/2019/02/sst-tropical-pacific-rotated.jpg
Is the cloud positive feedback to SST counter intuitive?
https://www.mdpi.com/2225-1154/6/3/62
https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2018GL078242
To my mind a century long more positive planetary energy imbalance can be inferred from these short term mechanistic observations. With implications for attribution of ocean heat and surface temperature changes to almost exclusively anthropogenic causes. And for the future evolution of climate as ENSO epochs tend to revert to a mean state. Although the latter seems far from guaranteed – even with mooted propitious solar winds or lunar tides.
Judith,
Your most enjoyable essay would be enhanced for this reader if you could assure us that the SST is now so much a part of the story on these eddies, that it has reached the status of a ‘given’.
For years I had imagined that a change of SST from (say) 300K to 301K lacks both the inherent energy content to make a difference, and to power the dynamic gradients that allow the energy to be moved to the critical place at the critical times.
Australian cyclone tracks have been depicted as lasting for days and a couple of thousand km over dry, parched land of undetermined surface temperature. But maybe the part of temperature in sustaining cyclones is not too related to the initiation of them. Geoff
http://www.geoffstuff.com/tracks.jpg
Blog Readers and Dr. Curry. Scott Adams has put out a challenge for the best 5 arguments Pro and Con regarding Climate Change. Would you muster your resources and take on the challenge? Here is an example and please challenge other bloggers to do the same.
Response to Scott Adams; The CO2isLife Top 5 Skeptical Arguments
https://co2islife.wordpress.com/2019/02/21/response-to-scott-adams-the-co2islife-top-5-skeptical-arguments/
We can then create a top 5 of the top 5 list
Imo, the White House is sort of the last bastion of climate denial.
Do you honestly believe spending trillions of dollars will stop the climate from changing?
You climate alarmists are the ultimate climate denial world champions. You have no clue to what causes natural climate change.
We elected Trump to fix this mess and he is working on it. The White House is our front line in this climate war. Fossil fuel and CO2 have only accomplished good in history and will only accomplish good in our future.
There is no actual data to prove CO2 has ever caused harm or that CO2 will ever cause harm. There is plenty of actual data that proves the benefits of more CO2.
Last bastion
Yes, and all those who actually engage in critical thinking. Admittedly, this is a diminishing skill. But we persevere.
Does the U.S. need a ‘presidential climate security committee’? – 02/22/2019
https://i.imgur.com/uOU7xtR.png
Earlier to that in that article JC said:
“Judith Curry, a former Georgia Tech climate scientist who has frequently been invited by Republicans to testify on climate science, because of her outspoken focus on the uncertainties, said the tone among climate skeptics in Washington seems to be shifting away from outright denial. Some anti-regulatory think tanks, she noted, are shutting down their climate-focused programs. Congressional hearings no longer center on yes-no climate debates.”
I was once sent a list of about 90 comments Trump made on climate change. Only 4 were actually defensible from a skeptic point of view. Unfortunately the debate on climate change has become so polarised down partisan lines that even if the skepticism is justified in the sense that it could be, the position is taken in political opposition not because of genuine knowledge or understanding. And the same goes the other way just as badly, with the added derision (also justifiable) of the term “denier” as if those taking a skepitcal position were the equivalent of anti-vaxers, flat-earths or even holocaust deniers. It’s an equivalent stupidity.
Over a longer term, it doesnt appear that hurricanes are increasing at all. https://www.gfdl.noaa.gov/historical-atlantic-hurricane-and-tropical-storm-records/
ScienceMag stupidly says Happer is “a prominent opponent of climate science”.
https://www.sciencemag.org/news/2019/02/retired-physicist-leading-new-trump-effort-question-climate-threat-security
The article is reprinted from E&E News. It systematically and stupidly equates alarmism with “climate science.” Skeptics have a good grasp of the actual science. What they object to is alarmism’s interpretation. Hence the scientific debate, which this article seems to claim does not exist.
The author is Scott Waldman at swaldman@eenews.net.
Feel free to correct him. I have already.
My favorite argument is that sea level has risen 120 meters over the last 20,000 years.
Of that, only 8 inches happened last century. So even if CO2 caused all 8 inches of SLR during the last century (probably not true), 8 inches is only 1.7% of the 120 meter rise – which was caused by totally natural phenomena.
How do we know we are not still rebounding from the LIA?
Why should we assume that natural phenomena have stopped and all of the warming is caused by human activity?
Why did it warm from 1905 to 1945? Wasn’t that at least in part natural and so how much of the 8 inches of SLR was caused by nature and how much by humans (in the last century)?
For all we know, we could be talking about 4 inches of SLR being caused by humans – which is only about .8% of the 120 meters since the last ice age.
To pretend that all further warming is 100% due to humans from now until 2100 (or 2200 or 2300) seems completely unwarranted to me.
Just one lay person’s opinion.
You Wrote:
“Attribution studies estimate that about a half (40–54%) of the global warming from 1901 to 1950 was forced by a combination of increasing greenhouse gases and natural forcing, offset to some extent by aerosols. Natural variability also made a large contribution. The exact contribution of each factor to large-scale warming remains uncertain.”
The Roman and Medieval warm periods were warmed into from out of colder periods without man-made CO2. If you are uncertain why that happened, how can you know that cause suddenly stopped and was replaced by man’s influence. You identified “Three modes of interannual to multi-decadal variability”. A change from several hundred warm years into several hundred cold years and then to several hundred warm years to several hundred cold years is multi-century variability. The warm periods are the times that sequestered ice in cold places builds up due to more evaporation of Polar Oceans and more snowfall until the ice advances and causes cold periods. The cold periods are the times that there is a lack of evaporation of frozen Polar Oceans and lack of snowfall and a depletion of the sequestered ice until it retreats and causes another warm period. This is recorded in the ice cores from Greenland and Antarctic. More ice extent causes more cooling from reflecting and thawing. Less ice extent causes less cooling from reflecting and thawing. The ice extent change is cause of temperature change and not result.
This buildup of sequestered ice is happening now. The volume and weight are countering the ice extent retreat out of the little ice age. Everyone talks about the ice retreat due to warming. Data is available to show that Piles of ice expand when there is more volume and weight and piles of ice retreat when there is less volume and weight.
People who do long range forecasts do understand that warmer oceans with more open Arctic in September do promote early snowfall on the Northern Hemisphere Continents that do influence the Jet Stream and cause the Polar Vortexes to promote more cold and snowfall on land in America and Europe and Asia and even reaches Africa some years.
Warm periods are natural, normal and necessary in the natural cycles of climate change. The major point is not that changing climate phases do not cause different weather results, because it really does. Warm periods have different weather events and conditions than cold periods. Man-made CO2 did not cause the differences, the different cycles of climate happened forever in history and we did not suddenly take over. Warm oceans do promote open Arctic and does promote more evaporation and more snowfall and does influence polar vortexes and weather events. To say what is happening with snowfall this years is not different, may even have some truth, but none that will help you. Daily disaster news rules. Stick to the fact that everything is inside the bounds of past history and we did not cause it. CO2 has not caused this warm period to be warmer than the Medieval or Roman Warm period, but all warm periods have some weather events that are different than events in cold periods.
The weather events have changed, some better some worse, but they have changed due to changing phases of the natural climate cycles. If you cannot prove everything is not worse, do not say and write it.
It is easy to prove natural climate has cycled in the past and everything is inside the bounds of past changes, we have data and history to support that. The data supports that there has been changes over the most recent century of “official recorded data” but there is no data to support that any of it has not happened before, other than CO2 which is much better, thanks for more to eat because of that.
Climate changes in natural cycles and Man has not suddenly taken over. The changes are natural, normal, necessary and unstoppable.
Open Arctic due to warm thawed oceans and warmer oceans around the Arctic does cause more evaporation and snowfall in the northern hemisphere, data and history prove this. Please quit saying it is not true. You lose every time there is a major storm across the country.
Ice sequestered on land increases when it is warmer and when oceans are thawed and there is more evaporation and snowfall.
Ice sequestered on land depletes when it is colder and when oceans are frozen and there is less evaporation and snowfall
This causes natural, normal and necessary warm and cold cycles that will never stop. The amount of ice and water that has taken part in each cycle has changed and that is why cycles now are different than cycles in the past.
More evaporation and forming of water and ice in clouds and more IR out does cool the Earth. It does not cool the earth all then, it creates ice that does more cooling later.
When oceans are warmer, Polar Oceans are thawed, and this process is used to create and sequester ice. Power in and energy dissipation is at a max but Earth is still warmest then.
When oceans are colder, Polar Oceans are frozen and this process is turned down and ice is allowed to dissipate. IR out is least then so lack of IR out fails to warm the earth. Earth is coldest then because sequestered ice has advanced and is doing the most cooling by reflecting and thawing.
This process is toggled up and down based the temperature that sea ice forms and thaws. This is the thermostat set point for the Polar Ice Cycles to increase or decrease Ice Sequestering in cold places. Tropical Climate also has more evaporation and precipitation when oceans are warmer and less when cooler, but the ice cycles do not play a part. Temperate Climates are influenced by both the Tropical Climate and Polar Ice Cycles. The ice cycles in the North and ice cycles in the South both influence temperature and sea level and they are sometimes in phase and sometimes out of phase, they run at their own internal natural frequencies that depend on the volume of water and ice that are part of each cycle.
Major Ice Ages shared a common ocean so the major cycles were coordinated by more snowfall when oceans were deep and warmer and by less snowfall when oceans were shallow and colder. Major Ice Ages were North Hemisphere Events. Antarctic went along because of the shared oceans. Ice ages are a time with more ice, earth and oceans are cold and oceans are low and cold and there is not much snowfall in cold times. Warm times are a time with less ice, earth and oceans are warm and oceans are high and warm and thawed. This is the time that more evaporation and IR out can produce ice and sequester it on land.
Little Ice Ages and Warm Periods are now multi-century cycles. Major Ice Ages and Warm Periods were multi-thousand-year cycles. No such change has happened with external forcing cycles, the difference is internal with the amount of ice and water that take part in the cycles.
There are internal natural cycles with their own natural frequencies. External forcing cycles resonate with these internal cycles differently at different times. Most people only look at correlations between temperature and external cycles. Models are built using those correlations. Theory and models remove ice when oceans are warmer and thawed. Data increases ice accumulation when oceans are warmer and thawed.
Yes it is quite obvious that ice is accumulating at a rate of knots. The iceman cometh.
There is even more snow in cold places in Hawaii this year
https://www.sfgate.com/weather/article/Hawaii-recorded-what-may-be-the-lowest-elevation-13607099.php
It snows more when oceans are warmer and more thawed and there is more ice sequestered that will build and cause a cold period later. This is Multi-Century climate cycles during their ice rebuilding time period.
The news is full of snow stories lately and another snow storm appears in the west as one ends in the east. Each of these storms are delivering snow that will thaw and disappear soon or as summer gets warmer. Some of these storms are delivering snow that will become sequestered ice that will increase volume and weight of ice sheets and glaciers that will advance after a few hundred years of more snowfall and it will get colder again like it does about every thousand years, give or take a few hundred, with lots of other correlations with other forcing.
What has happened will happen again. Study ice core data and history.
We are warm now. Earth will get colder again after a few hundred years of warmer with more snowfall and ice sequestering and finally more increase of ice extent.
In 1484 it was already known that climate change is always anthropogenic.
Summis desiderantes affectibus:
Many persons of both sexes, unmindful of their own salvation and straying from the Global Warming Faith, have abandoned themselves to devils, Big Oil and the Koch brothers, and by their incantations, spells, skeptical utterances, and other accursed charms and crafts, enormities and horrid offences, have … blasted the produce of the earth, the grapes of the vine, the fruits of the trees, nay, men and women, beasts of burthen, herd-beasts, as well as animals of other kinds, vineyards, orchards, meadows, pasture-land, corn, wheat, and all other cereals.
What are you saying? I do not understand which side, if any, you are supporting. Please explain.
What I’m implying is that attributing hurricanes and in fact practically all adverse or news-worthy weather events to human carbon emissions, has as much scientific support and basis as the attribution of such events 500 years ago to witchcraft.
“If oceans are getting warmer as a result of climate change, so the argument goes, surely hurricane activity must increase as a result, particularly hurricane intensity.”
–
There is an argument, many have seen it, that in a warmer world the same extra heat that allows more hurricanes to form actually inhibits their formation as the rate of change of increase in heat is perforce slower at higher temperatures.
I do not know what tho think.
It is a bit like the ridiculous argument that warmer wearpther causes more snow.
Still, it is out there.
It enables warmists to claim their two bob either way (each way bet for Americans).
Any thoughts on this conundrum?
Here is the BOM Australia region graph since 1970 showing severe and non severe cyclones. The trend goes down for both over that period.
And cyclones were much worse in the past and Dr Nott studies show the last super cyclone hit the NE Qld coast over 200 years ago. See Catalyst ABC super cyclones video.
http://www.bom.gov.au/cyclone/climatology/trends.shtml
Here is the ABC Catalyst story about Dr Nott’s study of OZ cyclones over the last 6,000 years.
http://www.abc.net.au/catalyst/stories/s382613.htm
Pingback: Weekly Climate and Energy News Roundup #349 |
Pingback: Unbewohnbare Erde: Ein Abraham a Sancta Clara d. Klimakatastrophe
Where is actual data of “pre-industrial” era rainfall that would make this claim serious?
“As a recent example, Patricola and Wehner (2018) used a high-resolution model to simulate 15 hurricane events from the global historical record. Simulations for each storm were conducted under current climate conditions versus the surface climate associated with pre-industrial conditions. They found that, relative to pre-industrial conditions, climate change has enhanced the average and extreme rainfall of hurricanes Katrina, Irma and Maria by 4%–9% and increased the probability of extreme rainfall rates, suggesting that climate change to date has already begun to increase tropical cyclone rainfall.
“The model used by Patricola and Wehner (2018) was driven by specified sea surface temperatures, and did not include coupling to the ocean. Lack of ocean coupling in the model can lead to tropical cyclones that are more intense and frequent compared to slab…”
Focus on the strong assertion ” climate change has enhanced the average and extreme rainfall of hurricanes Katrina, Irma and Maria by 4%–9% and increased the probability of extreme rainfall rates, suggesting that climate change to date has already begun to increase tropical cyclone rainfall.”
There is no data of detailed worth to verify this ‘before and after’ narrative. There is nothing but modeled assumptions and multiple attendant presumptions that these – and these alone – explain these modeled “experiments” results.
GIGO in climastrology, Judith? Such ought to be the real scientist’s warning label.
Judith: in many places, you refer to the impact of global warming on hurricanes – when it might be more appropriate to refer to the impact of rising GHGs on hurricanes. (I’m not sure whether this is a critical difference or pickiness from someone who is overly focused on “causation”.) For example, you conclude with:
“Models and theory suggest that hurricane intensity and rainfall should increase in a warming climate. There is no theory that predicts a change in the number of hurricanes or a change in hurricane tracks with warmer temperatures.”
Global warming is defined by rising surface temperature, SST over 70% of the planet. Hurricane intensity is not a simple function of rising SST. Rising absolute humidity, but not relative humidity, is a function of warming. (Rising relative humidity over the ocean is a function of slower atmospheric turnover.) For example, Walsh (2015) has a section discussing: “How does the role of changes in atmospheric carbon dioxide differ from the role played by SSTs in changing tropical cyclone characteristics in a warmer world?”
“decreases in midtropospheric vertical velocity are generally larger for the 2CO2 experiments than for the 2K experiments”
“First, Held and Zhao (2011) showed that one of the largest differences between the results of the 2K and 2CO2 experiments conducted for that paper was that PI increased in the 2K experiments but decreased in the 2CO2 experiment, owing to the relative changes in surface and upper-tropospheric temperatures in the two cases. In addition, directly simulated intense tropical cyclone (hurricane) numbers decrease more as a fraction of their total numbers in the 2CO2 experiment than they did in the 2K experiment, consistent with the PI results.”
https://journals.ametsoc.org/doi/pdf/10.1175/BAMS-D-13-00242.1
I had criticized your previous post for focusing on detection (changes over time) rather than causation (change with rising SST, which you called attribution). After reading Walsh, the variable on the x-axis of any causation graph probably should be CO2 or CO2 equivalents, not time or SST. None of your graphs have SST or CO2 on the x-axis. If you constructed one, the period of accurate global hurricane records (since 1990?) would cover a respectable 60 ppm increase in CO2, half of the total change. The period of accurate Atlantic records (since 1970) increases the dynamic range by only 20 ppm. Unfortunately, natural variability is problematic.
(My perspective is shaped by having had the luxury of running well-controlled laboratory experiments, where I determine what variable will be plotted on the x-axis. Climate scientists have a much tougher job.)
well one of my main conclusions is that SST (and particularly CO2) don’t really have all that much to do with hurricanes
Let’s hypothesize that evaporation rises 7%/K. That is an additional 5.6 W/m2 of latent heat per K of surface warming entering the atmosphere (-5.6 W/m2/K). The change in net LWR radiation (OLR-DLR) with surface warming is negligible compared to this much latent heat. However, if ECS is 3.6 or 1.8 K, only a net -1 or -2 W/m2/K is crossing the TOA. The change in the surface energy balance with warming (in W/m2/K) must be equal to the change with warming at the TOA. Every +1 W/m2/K of positive SWR feedback is a 1%/K reduction in reflection of SWR. A change much bigger than 1%/K seems absurd when thinking about glacial periods, so the gap between -1 or -2 W/m2/K and -5.6 W/m2/K probably can’t closed by positive SWR cloud feedback. Therefore our hypothesis that evaporation and precipitation increase 7%/K must be wrong (or ECS is absurdly low or cloud SWR feedback is absurdly positive).
Therefore, when CO2 slows radiative cooling of the upper troposphere, convective turnover of the atmosphere must slow, causing relative humidity over the oceans to rise. Higher relative humidity slows evaporation, causing precipitation to increase less than 7%/K. Slower overturning might also be accompanied by slower trade winds, which would further reduce the rate of evaporation. Positive feedback and high ECS appear to be intimately linked energetically to a slowing of convective turnover of the atmosphere and rising humidity of the ocean. Convective turnover and humidity play an essential role in hurricane genesis.
So perhaps 2 K warming of SSTs caused by rising GHGs is different than a 2 K warming artificially imposed in climate models. Walsh (2015) devotes a whole section to the difference in hurricanes modeled in “2K” and “2CO2”. Now that I have opened my big mouth (and inserted foot), I realize that I didn’t understand the difference between these modeling experiments. I just assumed that it was related to the changes in convection and humidity caused by rising GHGs rather than changes produced by rising temperature alone.
In general, I agree with your reply that SST’s don’t have much to do with hurricanes, but the first sentence of your conclusion (which I quoted) reads differently to me. I’m personally interested in seeing more people doing what Pat Michaels did: correlating hurricane wind speed to the current SST of the ocean below and other factors. (Not the disturbed ocean directly below; the adjacent ocean “feeding” humidity into the wind being swept into the hurricane.) Michaels didn’t look at the other factors: humidity, the outlet temperature at the top of the hurricane, the temperature difference between SST and outlet, winds in the upper atmosphere to dissipate all of the released latent heat released, wind sheer, the rate of hurricane intensification. Does being over a 29 K ocean double the probability of intensification from Category 2 to Category 3 in the next 24 hours compared with being over 27 K ocean? If the probability of intensification doesn’t increase with SST, how can rising SSTs be responsible for more major hurricanes? Tony Merlis’s paper seems to collect some of this information, but every graph shows change over time.
In theory, climate models and weather prediction models contain all of the information I’d like to know and take it into account. However, they don’t perform a multiple regression to determine which factors (if any) are consistently important. And one would like to prove that the important factors in the real world are being reproduced by the model.
Pingback: Hurrikane und Klimawandel 2: Zuordnung – EIKE – Europäisches Institut für Klima & Energie