Earlier start of North Atlantic hurricane season with warming oceans | Panda Anku

Historical data of tropical cyclones

All historical data for Atlantic tropical cyclones are from the HURDAT2 dataset2 for 1900-2020. For historical CONUS tropical storm observation and alert information, see the National Hurricane Center’s Tropical Cyclone Reports for each April 1-May 31 TC that formed in 2012-2020 and for the CONUS monitors or alerts were required31,32,33,34,35,36,37.

ACE is a built-in metric that takes into account the frequency, intensity and duration of storms. AS38 Calculations are performed using the six-hour HURDAT2 intensity data, calculating the ACE contribution for each six-hour maximum sustained wind v in knots:

$${ACE}=frac{{v}^{2}}{10000}$$

(1)

and all entries for which a TC or subtropical cyclone (STC) has v >= 34 knots summed to give annual totals. Extratropical cyclone dates are not included.

USACE calculations are performed as described in the methodology of Ref.1. 6. Based on the HURDAT2 data, we consulted the Atlantic Oceanographic and Meteorological Laboratory’s hurricane landfall list39 to identify systems that affected the CONUS between 1966 and 1981 when HURDAT2 does not explicitly include landing points. We then used existing range maps to qualitatively estimate a landfall location and an intensity data point to the nearest hour, which we manually added to the HURDAT2 dataset for these storms. These position and intensity data were linearly interpolated to a time resolution of one hour. We use this hourly dataset to calculate the hourly ACE from maximum sustained winds v in knots by doing:

$${AC}{E}_{h}=frac{{v}^{2}}{60000}$$

(2)

To obtain an annual time series from USACE, we applied a spatial landmask to this hourly ACE metric to exclude all TC positions further than 0.5° from any part of the CONUS. We chose the 0.5° buffer to match the typical radius of maximum winds of a mature TC and to account for possible small observation errors, especially at the beginning of the recording40. For the purpose of counting, we consider a “landing” as the center of circulation of a TC or STC with v >= 34 knots passing within 0.5° of CONUS for each interpolated time step of one hour. A single TC or STC may make multiple CONUS landings, but may only make a maximum of one landing in each of the coastal regions shown in Fig. 4a of Ref. 4a. 6. To find the annual values ​​of USACE, we summed this series over the CONUS and in time over each hour of the year. All time series, distributions, and counts in this document assume that the annual cycle of TC activity begins on March 1 of each year at 0000 UTC and ends on February 28 or 29 of the following year at 2359 UTC. This choice is made to best correlate with the mean annual minimum of North Atlantic SSTs and hence TC activity in January and February, such as Hurricane Alex in 201641 is considered part of the annual cycle of the previous year’s North Atlantic TC activity. January-February TC and STC formations account for less than 0.5% of total annual formations over the period 1950-2020.

Quantile regression and trend calculations

The trends of the date of the first named Atlantic storm were calculated from ordinary least squares regression of the annual series of the number of six-hour periods between 0000 UTC March 1 and the first six-hour HURDAT2 TC or STC with sustained winds >= 34 knots year-on-year, subsequently converted back to a rate of days year−1. This calculation was repeated to exclude any TC or STC with less than 8 six-hour HURDAT2 entries with maximum sustained winds >= 34 knots to test the impact of short-lived systems on the overall trend. Trends in the first CONUS impact date were obtained from the ordinary least squares regression of the annual series of the number of one-hour periods between 0000 UTC on March 1 and the first one-hour interpolated HURDAT2 TC or STC entry with sustained winds >=34 Knots calculated to within 0.5° or less of CONUS over the year, calculated back at a rate of days per year−1. The significance of all trends was assessed using a Mann-Kendall test42.43. The fractions of TC activity occurring within the current official definition of hurricane season are reported as a percentage of all TC and STC initial formations from March 1 through March 28/29. February with sustained winds >= 34 knots, CONUS TC and STC landings with sustained winds >= 34 knots indicated, called storm days, ACE and USACE occurring between 0000 UTC Jun 1 and 2359 UTC Nov 30. These data are computed in 50-year lagging average windows to account for multidecadal variability in Atlantic TC activity13.

Trends in percentile threshold data for ACE (both a climatology that includes short-lived TCs and one that excludes them according to the criteria above) and USACE are determined using quantile regression44. Quantiles divide the threshold dates at which the specified percentage of total annual ACE and USACE is achieved, here from 1% to 99% in 1% intervals, into equal subsets. Quantile regression applies ordinary least squares regression to conditional quantiles of the ACE and USACE response variables, taken at regular intervals from the cumulative distribution of ACE and USACE over 1979-2020 and 1900-2020, respectively, for regressions against the year to be carried out. The sensitivity of the calculated trend to the choice of time series length is tested by repeating this quantile regression method for the starting years of the cumulative distribution functions for ACE from 1950–1990 and USACE from 1900–1990 with a fixed ending year of 2020.

Synoptic environment sensitivity tests for quantile regression

The dependence of the ACE and USACE percentile threshold data on synoptic environmental factors is tested by performing quantile regressions of these response variables against GPI, SST, 200 T, RH, and VWS. These regressions are performed by the ACE and USACE cumulative distribution functions over 1979–2020, due to the superior quality of global reanalyses from 1979 onwards. The SST time series used in the quantile regressions are computed from the April and May monthly mean ERSSTv522 SST fields averaged over 10–36°N, 100–70°W, excluding portions of this box over the Pacific. The spatial resolution of ERSSTv5 is 2°. RH, 200T and VWS time series are calculated from the monthly mean of April and May ERA521 Fields. The primary GPI time series used in this study23 is calculated from the monthly means of the ERA5 fields from April and May. The GPI value is calculated at each 0.25° grid point in this field according to the equation23.25:

$${GPI}={abs}({10}^{5}{{eta }})^{2}{* left(frac{{{{{{{rm{RH}}}}} }}{50}right)}^{3}{* left(frac{{{{{{{rm{PI}}}}}}}{70}right)}^{3}* { (1+0.1{{{{{rm{VWS}}}}}})}^{-2}$$

(3)

with η the 850 hPa absolute vorticity, RH the 700 hPa RH, PI the potential intensity, a theoretical TC intensity maximum at a given SST and a specified atmospheric column25.45and VWS is the vertical wind shear between 850 and 250 hPa. The GPI statistic was developed using a fitting procedure of these variables in the NCEP/NCAR reanalysis46 for seasonal and spatial climatological recording of global cyclogenesis events. These grid point GPI values ​​are then averaged separately for April and May over 10-36°N, 100-70°W, and the two months are then averaged to obtain the GPI time series used in the quantile regressions.

An alternate form of GPI24 was used at several points in this study for comparison purposes using the same data sources described above. This construction of GPI has the form:

$${GPI}=|{{{{{rm{eta }}}}}}|^{3},*, {{chi }^{-frac{4}{3}} ,*,left(right. {{{{{rm{MAX}}}}}}(({{{{{rm{PI}}}}}}-35m{s}^{ -1}),0)}^{2}* {(25{{ms}}^{-1}+{{{{{rm{VWS}}}}}})}^{-4}$ $

(4)

with η, PI and VWS as above, and

$$chi =frac{{s}_{b}-{s}_{m}}{{s}_{0}^{* }-{s}_{b}}$$

(5)

with χ a measure of the wet entropy deficit in the middle troposphere and sbsm, and s*0 Humidity tropy of the boundary layer and the middle troposphere and saturation humidity tropy of the sea surface24.

TC emergence SST threshold and SST trends

Since no daily SST values ​​are available from the ERSSTv5 dataset, the spatial coverage of the 26.5 °C threshold is calculated from the daily mean ERA521 SST fields that have a spatial resolution of 0.25°. ERA5 SSTs are from several different analyzes over the period 1950-2020, including HadISST2.1.0.0 before 1961, HadISST2.1.1.0 between 1961 and mid-200747and OSTIA since mid-200748. The native temporal resolutions of these analyzes are one month, five days, and one day, respectively. However, since SSTs slowly change over time, using these datasets scaled down to a uniform daily resolution is unlikely to introduce significant observational bias into the SST threshold analysis. The fraction of the field 10-36°N, 100-70°W for which this criterion is met is calculated by applying a landmask and excluding the Pacific Ocean and the number of daily grid points exceeding 26.5°C , dividing by the total number of non-land, non-Pacific grid points. Trends in ERSSTv5 monthly mean SST values ​​are computed from ordinary year-on-year least squares regressions. Box or global average SST trends are calculated after excluding land grid points. Supplementary Fig. 7 was the only case where ERA5 SST scores were used as the source for SST scores instead of ERSSTv5 monthly means.

Optimized limits for hurricane season

The objective start and end dates for the Atlantic hurricane season are calculated from 50-year average time series of named storm formations, CONUS storm impacts, Atlantic named storm days, Atlantic ACE, and USACE using a 15-day intraseasonal smoothing filter normalized for total activity within of the window. This results in a daily time series of the smoothed percentage of total TC activity recorded each day from March 1st to March 28th/29th. February occurs within the 50-year window. A middle-out algorithm begins at the climatological peak of Atlantic TC activity in mid-September and adds one day at a time to the target season until the period containing the fewest consecutive days for a given percentage activity threshold is reached. The tested thresholds are 95%, 97%, and 99% of total activity for each of the TC metrics.

Leave a Comment