Entity

Time filter

Source Type

Asheville, NC, United States

News Article
Site: news.yahoo.com

NOAA's GOES-West satellite captured this photo of the winter storm on Jan. 21, 2016, at 10 a.m. EST. More A massive winter storm that is expected to bring snow and ice to the eastern United States in the next 48 hours dwarfs the central part of the country in a new satellite image. The National Oceanic and Atmospheric Administration's GOES-West satellite spotted this cloudy view of the large storm near the Gulf Coast today (Jan. 21) at 10 a.m. EST. (The satellite photo also spied a separate storm swirling over the Pacific Ocean.) The NASA-NOAA Suomi NPP satellite captured another view of the looming winter storm yesterday (Jan. 20) at 2:30 p.m. EST, showing clouds and snow cover stretching from northern Texas into the Great Lakes states. The low-pressure system, which originated in the Eastern Pacific, is headed toward the mid-Atlantic states. The National Weather Service is warning of a potentially "crippling" snowstorm in those areas tomorrow and Saturday. [VIDEO: Watch the Developing Storm from Space] Much of the East, from as far south as northern Georgia to as far north as New York City, is under weather alert, ranging from winter storm warnings in eastern Tennessee, Kentucky and western North Carolina to a blizzard watch in Washington, D.C, Baltimore and up the I-95 corridor through New Jersey. "At this moment, this still looks like it's going to be northern Virginia and Maryland's most memorable snow," said David Robinson, the state climatologist of New Jersey. "A question remains just where exactly the heaviest snowfall totals will fall, just where the strongest winds will be and where the coastal flooding will be worst." The storm will track across the Tennessee Valley tonight and Friday, ultimately hitting the mid-Atlantic states Saturday, with lessening snow expected by Sunday. Storms like the one bearing down on the eastern United States typically have a band of heavy snowfall about 100 miles (161 kilometers) wide, Robinson told Live Science. North of this band, precipitation peters out, leading to low overall totals. South of this band, temperatures are warm enough that snow turns to rain. "This is a multifaceted nor'easter, as they all tend to be, laden with strong wind and impacts on the coast and heavy precipitation of one form or another," Robinson said. El Niño's influence may also be at play, Robinson said. The storm "fits the pattern of volatility that we often see in weather, really across the country, during an El Niño winter," he said. The winter of 2015-2016 has so far been marked by an active storm pattern in the western United States, with unseasonably warm temperatures and unusual midwinter thunderstorms in the South. In late December and early January, the Mississippi River flooded to levels not seen since 1993, thanks to heavy snows and rainfall in the western and central parts of the country. A rare December tornado outbreak from Indiana to Alabama killed 13 people between Dec. 23 and Dec. 25. Meanwhile, warm temperatures made Christmas a balmy holiday for many on the East Coast. New York City experienced its warmest Christmas on record, with temperatures reaching 66 degrees Fahrenheit (19 degrees Celsius). On Christmas Eve, the mercury hit 72 degrees F (22 degrees C) at LaGuardia Airport in New York, breaking a 76-year record. The previous high temperature that day, according to NOAA's National Centers for Environmental Information, was 63 degrees F (17 degrees C). If volatility is a feature of El Niño winters in the United States, so is an active East Coast storm track, Robinson said. "We have seen coastal storms aplenty during El Niño winters," he said. Those storms don't always translate to snow, though, he added. Some mid-Atlantic El Niño years have been virtually snow-free, while others have seen large snowstorms. Regardless of snow totals, the incoming storm is likely to be a major event, meteorologists say. Forecasts predict water levels 3 to 4 feet higher than normal along the New Jersey coast down into Delaware, Robinson said, and winds could gust more than 60 mph (96.5 km/h). Washington, D.C., and the surrounding area is under a blizzard warning, with about 24 inches (60 centimeters) of snow expected for the capital city and 24 to 30 inches (60 to 76 cm) in the western suburbs, according to the National Weather Service. Gusty winds, ice and up to a foot (30 cm) of snow are expected in Kentucky. Detailed regional weather advisories and forecaster discussions can be found at http://www.weather.gov/. Follow Stephanie Pappas on Twitter and Google+. Follow us @livescience, Facebook & Google+. Original article on Live Science. Copyright 2016 LiveScience, a Purch company. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.


News Article | February 6, 2016
Site: www.theenergycollective.com

The latest State of the Climate report from the National Oceanic and Atmospheric Administration (NOAA) continues with its broken record of broken records. The NOAA report states that not only was 2015 the warmest year on record, it broke the record by the widest margin ever. Average global land and ocean surface temperature was 1.6° F (0.90° C) above the 20th century average, beating the 2014 record of 0.29° F (0.16° C). 2015 marks the fourth year this century that global temperatures have broken the 136-year record. Global land surface temperature in 2015 was 2.39° F (1.33° C) above the 20th century average, beating the record set in 2007 by 0.45° F (0.35° C), again the widest margin for annual average land surface temperature. For the oceans, the global surface temperature in 2015 was 1.33° F (0.74°) above the 20th century average, beating the record set last year. Scientists consider not only continuing broken temperature record, but the margin by which they were broken in 2015 as significant. “A lot of times, you actually look at these numbers, when you break a record, you break it by a few hundredths of a degree,” Thomas Karl, director of NOAA’s National Centers for Environmental Information, said in the Washington Post. “But this record, we literally smashed. It was over a quarter of a degree Fahrenheit, and that’s a lot for the global temperature.” For December, average global land and sea surface temperature was 2° F (1.11° C) above the 20th century average, the highest in the 1880-2015 record, beating the previous record warmth set in 2014 by 0.52° F (0.29° C). December 2015 was also the highest departure of all months in the historical record above the 20th century average, for the first time reaching 2° F above any previous month’s average. Certainly some of the record-breaking heat in 2015 is attributable to the powerful El Niño that has driven enormous heat in the Pacific Ocean, but in contrast to the last big El Niño in 1998, 2015 was considerably hotter and the warmth spread across almost all regions of the globe, with one month after another breaking records. 1998, a banner year for climate “skeptics” claiming that’s when global warming “stopped,” now ranks “5th or 6th.” Well below many subsequent record-breaking years. “It’s breaking the record because we also have this unusually strong El Niño, but at the same time we know the ocean is now absorbing two times more heat than around the last time we had a big El Niño, which is quite a while ago,” saidTexas Tech University climate scientist Katharine Hayhoe. NASA and NOAA keep separate temperature datasets which don’t always perfectly agree. Gavin Schmidt, NASA’s director at the Goddard Institute for Space Studies, says the two datasets show “relatively little disagreement this year.” Schmidt also concurred that the El Niño does not alone account for the exceptional record-breaking warmth throughout 2015.


News Article | February 6, 2016
Site: www.theenergycollective.com

The latest State of the Climate report from the National Oceanic and Atmospheric Administration (NOAA) continues with its broken record of broken records. The NOAA report states that not only was 2015 the warmest year on record, it broke the record by the widest margin ever. Average global land and ocean surface temperature was 1.6° F (0.90° C) above the 20th century average, beating the 2014 record of 0.29° F (0.16° C). 2015 marks the fourth year this century that global temperatures have broken the 136-year record. Global land surface temperature in 2015 was 2.39° F (1.33° C) above the 20th century average, beating the record set in 2007 by 0.45° F (0.35° C), again the widest margin for annual average land surface temperature. For the oceans, the global surface temperature in 2015 was 1.33° F (0.74°) above the 20th century average, beating the record set last year. Scientists consider not only continuing broken temperature record, but the margin by which they were broken in 2015 as significant. “A lot of times, you actually look at these numbers, when you break a record, you break it by a few hundredths of a degree,” Thomas Karl, director of NOAA’s National Centers for Environmental Information, said in the Washington Post. “But this record, we literally smashed. It was over a quarter of a degree Fahrenheit, and that’s a lot for the global temperature.” For December, average global land and sea surface temperature was 2° F (1.11° C) above the 20th century average, the highest in the 1880-2015 record, beating the previous record warmth set in 2014 by 0.52° F (0.29° C). December 2015 was also the highest departure of all months in the historical record above the 20th century average, for the first time reaching 2° F above any previous month’s average. Certainly some of the record-breaking heat in 2015 is attributable to the powerful El Niño that has driven enormous heat in the Pacific Ocean, but in contrast to the last big El Niño in 1998, 2015 was considerably hotter and the warmth spread across almost all regions of the globe, with one month after another breaking records. 1998, a banner year for climate “skeptics” claiming that’s when global warming “stopped,” now ranks “5th or 6th.” Well below many subsequent record-breaking years. “It’s breaking the record because we also have this unusually strong El Niño, but at the same time we know the ocean is now absorbing two times more heat than around the last time we had a big El Niño, which is quite a while ago,” saidTexas Tech University climate scientist Katharine Hayhoe. NASA and NOAA keep separate temperature datasets which don’t always perfectly agree. Gavin Schmidt, NASA’s director at the Goddard Institute for Space Studies, says the two datasets show “relatively little disagreement this year.” Schmidt also concurred that the El Niño does not alone account for the exceptional record-breaking warmth throughout 2015.


We use the newly developed SST reconstructions and palaeoclimate databases of the PAGES Ocean2k working group. For full details of the data selection criteria and in-depth discussions of the marine datasets and reconstructions, see refs 14 and 15. The high-resolution (annual or better) component of the Ocean2k dataset consists of 57 coral records (including multiple sensors for some records; Extended Data Table 1), which were used to reconstruct regional mean SST histories for different sectors of the tropical oceans14. The SST reconstructions for the tropical Indian, western Pacific and western Atlantic regions are statistically robust over most of the past 400 years14. The same reconstruction method applied to the tropical eastern Pacific region yielded poorer statistics and a twentieth-century SST trend that is stronger than suggested by instrumental records. This was attributed to nonlinear hydrologic effects in the eastern tropical Pacific14; hence, we exclude this regional reconstruction from the current study. We use the ‘best’ reconstruction for each region, as identified in ref. 14 on the basis of validation statistics from the ensemble of reconstructions produced. From the Ocean2k low-resolution database15, we use a subset of 21 marine records that are suitable for the recent temperature trend analysis carried out in this study (Extended Data Fig. 7a). These 21 records meet additional criteria of having strong chronological control through 210Pb profiling or counting of annual sediment layers or coral growth bands, as well as an average sample resolution of 25 years or better; we refer to these as moderate-resolution records. The sense of the proxy-temperature response of the high- and moderate-resolution ocean palaeoclimate records is based on known physical relationships for the incorporation of geochemical and biological tracers into these records or on individually assessed and published temperature–growth relationships. We compare the ocean records to the continental-scale temperature reconstructions and palaeoclimate databases developed in phase 1 of the PAGES 2k project4, including the updated Arctic v1.1.1 reconstruction and database16. For the North America region we use the tree-ring-based reconstruction, which has decadal resolution, rather than the lower-resolution pollen-derived temperature reconstruction. To avoid record duplication, all marine records were removed from the PAGES 2k continental database before site-level trend analysis. We also exclude the four instrumental records in the South American 2k database from our site-level data assessment so that all information is derived solely from palaeoclimate archives. At the level of regional temperature reconstructions, a small degree of overlap exists in the contributing data used for some of the previously published reconstructions. Specifically, 9 of the 28 records used for the Australasia 2k temperature reconstruction are also used for the tropical western Pacific SST reconstruction, and 1 is used for the tropical Indian Ocean SST reconstruction. To avoid any bias introduced by data overlap, we use a terrestrial-only Australasia 2k reconstruction that was produced using the same methodology as the original reconstruction4, but excludes any of the marine geochemistry records that are used in the Ocean2k high-resolution reconstructions. Details of this terrestrial-only reconstruction, and the reconstruction data and statistics, accompany this paper as Supplementary Data 1. The terrestrial-only reconstruction demonstrates close agreement with the original Australasia 2k reconstruction and none of the interpretations presented here is altered by using the original reconstruction instead of the terrestrial-only version. Matlab data structures containing the site-level proxy data and regional reconstructions used in this study are archived with the National Centers for Environmental Information (NCEI) World Data Service for Paleoclimatology at http://www.ncdc.noaa.gov/paleo/study/20083. The analyses performed in this study use annually resolved, unsmoothed input data. For the moderately resolved marine records (resolution given in Extended Data Fig. 7a) and the North America reconstruction (decadal resolution), pseudo-annual data was produced by performing a nearest-neighbour interpolation to produce stepped datasets that continued values across the entire sampling interval that they represent. Chronological uncertainty in the palaeoclimate records and reconstructions is extremely low for the industrial era. Annual layer chronologies would be expected to be known to within ±2–3 yr back to ad 1800, using conservative estimates and not taking into account the 1815 Tambora eruption that left an unambiguous fixed-time marker in many palaeoclimate archives. Therefore, chronological uncertainty in the industrial era is well within the level of interpretability of our estimates for the onset of sustained, significant warming based on the change-point detection method (see Methods section ‘Change-point method testing’). The area weightings used to calculate the average tropical ocean temperature histories in Fig. 1d were based on the surface area of the target reconstruction regions on an Earth ellipsoid. These areas are: Indian Ocean, 25.5 × 106 km2; western Pacific, 26.9 × 106 km2; western Atlantic, 5.1 × 106 km2. The areas of the terrestrial reconstruction regions4 (Fig. 1c) are: Arctic, 34.4 × 106 km2; Europe, 13.0 × 106 km2; Asia, 31.1 × 106 km2; North America, 12.5 × 106 km2; Australasia, 37.9 × 106 km2; South America, 20.0 × 106 km2; Antarctica, 34.4 × 106 km2. The seasonality of the temperature signal captured by the reconstructions differs between regions, owing to the availability of site-level palaeoclimate records in each region, some of which capture climate information related to only a specific season (for example, a summer growing season for trees). Detailed discussion on seasonality can be found in refs 4, 14 and 15. To summarize: the tropical ocean reconstructions represent April–March (tropical year) annual averages; the Arctic, North America and Antarctic reconstructions represent annual averages; the Australasia reconstruction represents a September–February half-year (warm season) average; and the Europe, Asia and South America reconstructions represent local summer averages. We do not expect the seasonal differences between the regional reconstructions to affect our interpretations of the onset of sustained, significant warming between different regions. Change-point analysis of climate model simulations produces near-identical regional warming onsets when data are compiled as annual averages across all regions, and as season-specific averages that match the reported seasonality for each corresponding palaeoclimate reconstruction (Extended Data Fig. 4). The effect that uncertainty in the regional temperature reconstructions may have on estimates of the onset of industrial-era warming was assessed using reconstruction ensembles (Extended Data Fig. 1a). Reconstruction ensembles are available for South America, Australasia and the three tropical ocean regions, and were calculated as part of the original reconstruction process by using different methodological choices based on the proxy network, the temperature calibration interval and/or target dataset. We also extend this analysis to reconstruction ensembles available for Northern Hemisphere3 and Southern Hemisphere5 average temperature as an additional test of the apparent Southern Hemisphere delay in the onset of industrial-era warming (Extended Data Fig. 1a, Supplementary Fig. 1). Uncertainty in the onset of industrial-era warming related to reconstruction uncertainty has a 5%−95% range of within −3 yr to +25 yr for the three tropical ocean regions (which each have an average ±2σ range across reconstruction members of approximately 0.1 °C during the nineteenth and twentieth centuries). The Australasia ensembles have an average ±2σ range of 0.4 °C during the nineteenth and twentieth centuries, and there is a −35 yr to +46 yr range (5%−95%) in onset estimates determined across this ensemble. The South America ensembles have an average ±2σ range of 0.6 °C during the nineteenth and twentieth centuries, and a −16 yr to +31 yr range (5%−95%) in the onset of industrial-era warming (Extended Data Fig. 1a). As a first-order estimate, the ranges of onset timings obtained from regions for which reconstruction ensembles are available may provide guidance on how reconstruction uncertainty affects other regions. The regional reconstructions for the Arctic and Europe have a similar magnitude of uncertainty during the nineteenth and twentieth centuries to the South America reconstruction (average ±2σ of approximately 0.6 °C). Therefore, uncertainty in the onset of industrial-era warming related to regional reconstruction quality may be similar between these regions. Reconstruction uncertainty is higher for Asia (approximately 0.9 °C for ±2 root-mean-square error) and Antarctica (approximately 1.2 °C for ±2 standard error), and so we may expect a larger uncertainty range in onset estimates related to reconstruction quality for these regions. Reconstruction uncertainty for North America is based on decadal-resolution data, and so is not directly comparable to the annual resolution of the other regional reconstructions. No statistical methods were used to predetermine sample size. We compare the regional palaeoclimate reconstructions to a multi-model ensemble of transient last-millennium simulations from ad 850 to ad 185051 (Extended Data Table 2), completed as part of the Fifth Coupled Model Intercomparison Project (CMIP5). Historical simulations from the same ensemble were used to extend the model output from ad 1850 to ad 2005. The CMIP5 last-millennium and historical experiments use transient radiative forcings that include orbital, solar, volcanic, greenhouse and ozone parameters, as well as land-use changes17, 52. All data were accessed from the Earth System Grid Federation, with the exception of the historical portions of the HadCM3 simulation (provided by A. Schurer) and the FGOALS-s2 simulation (provided by T. Zhou and W. Man). Multiple simulations of the last-millennium run with LOVECLIM18 were used to examine climate responses to single radiative forcing scenarios (three ensemble members each) and to assess intra-model variability in full forcing simulations (ten ensemble members). We also examine multiple last-millennium simulations of the CSIRO Mk3L coupled climate model run with progressive addition of radiative forcings (three ensemble members each)21. The climate response to greenhouse gas forcing alone was further assessed using single forcing experiments of the HadCM320 (four ensemble members) and NCAR CESM119 (three ensemble members) models. For each of the models we examine surface air temperature (tas field in CMIP5 output files) averaged over the PAGES 2k continental palaeoclimate reconstruction regions, and sea surface temperature (tos field in CMIP5 output files) averaged over the PAGES Ocean2k tropical ocean reconstruction regions. Monthly resolution model output was used to generate annual or season-specific averages that correspond to each palaeoclimate reconstruction target region. Because we use the simulations to examine change-points rather than the magnitude of trends, we do not apply any drift corrections to the model output. Change-point analysis was performed on individual model runs and then compiled (Fig. 3, Supplementary Fig. 2, Extended Data Fig. 8), rather than on ensemble averages, to avoid loss of internal variability in the model data. Matlab data structures containing the model output (compiled as regional time series with annual-average and season-specific resolution) used in this study are archived with the NCEI World Data Service for Paleoclimatology at http://www.ncdc.noaa.gov/paleo/study/20083. We analyse the trends in the PAGES 2k continental and tropical ocean temperature reconstructions (Fig. 2a, Extended Data Fig. 1a), in the site-level terrestrial and marine palaeoclimate databases (Fig. 4, Extended Data Figs 6, 7) and in model simulations (Fig. 3, Extended Data Figs 2, 8, Supplementary Figs 2, 3) using the SiZer (SIgnificant ZERo crossings of derivatives) method24. SiZer determines the sign and significance of trends in time series data across different levels of smoothing using a Gaussian kernel filter. Following the method used in refs 14 and 23, we assess climate change-points from SiZer output by determining the median year of initiation for the most recent significant (P < 0.1) and sustained trends across smoothing bandwidths spanning all integer years in the range 15–50 yr. This range of smoothing levels is designed to reduce the influence of interannual-to-decadal climate variability on the detection of a sustained trend, while avoiding shifting the true change-point time if the smoothing window is too long. We assess trends in the time series since ad 1500, and the onset of the most recent significant trend is classified only if the sign of the trend persists through to the most recent end of the record (that is, a sustained trend). Extended Data Fig. 2a shows the SiZer data used to assess the median initiation point for recent significant warming trends across the continents and tropical oceans in palaeoclimate reconstructions (Fig. 2a). Supplementary Fig. 1 shows the SiZer data for multi-model palaeoclimate simulations (Fig. 3a). Change-points determined by the SiZer method were tested on a set of synthetic time series with known warming onset (Extended Data Fig. 3). We also compare SiZer estimates for the synthetic warming onset with change-points determined using linear change-point methods53, 54. The synthetic time series were designed to test the performance of change-point detection methods across different forms of long-term trends, in the presence of volcanic-style cooling events, and for varying magnitudes and redness (lag-1 autocorrelation) of variability superimposed upon the trend. Each test was carried out across 1,000-member ensembles to generate a distribution of change-point estimates for each method. Linear change-point detection methods best capture the change-point in synthetic time series when the long-term trend is derived from two straight lines (Extended Data Fig. 3a, series i, ii). However, the SiZer method is more adaptable than linear change-point methods in detecting the true change-point in time series in which the long-term trend is a curve rather than a straight line (Extended Data Fig. 3a, series iii). This is expected to be advantageous for detecting the initial thermodynamic response to increases in atmospheric greenhouse gas levels, which have an accelerating trajectory during the industrial era (Fig. 3d). Previous research has concluded that, despite the complexity of the climate system, there is a near-linear relationship between global radiative forcing changes and the climate response10, 55. Hence, we would expect industrial era climate trends to be better approximated by a curve than by a simple straight line. The addition of synthetic volcanic-style cooling events at the time of, and before, the change-point in synthetic series causes only small deviations in SiZer estimates of the climate change-point away from the true synthetic change-point (Extended Data Fig. 3a, series iv–vi). In our tests with a curved (accelerating) warming trend and volcanic-style cooling events centred at 0 yr, −25 yr and −50 yr relative to the onset of the warming trend, the SiZer method returns median times for the warming onset at −1 yr, −9 yr and −13 yr, respectively. This gives us confidence that large volcanic eruptions during the early nineteenth century are not likely to have substantially skewed the detection of the onset of industrial-era warming trends assessed using the SiZer method. Finally, we test the sensitivity of change-point detection methods to the addition of varying climate variability (autoregressive AR(1) noise) superimposed on the same long-term warming trend. As the magnitude of climate variability increases, the detection of change-points becomes progressively later using the SiZer method (Extended Data Fig. 3b). As a result, climate time series from regions with large interannual-to-multi-decadal variability may have delayed detection of the onset of industrial-era warming relative to regions with small variability, unless the magnitude of the underlying warming trend is also larger in these regions. In our testing, different levels of lag-1 autocorrelation for the AR(1) noise added to an underlying trend does not alter the median estimate for the onset of warming; however, the range of onset estimates about this median becomes greater as autocorrelation increases (Extended Data Fig. 3c). Across all of the tests examined here, the linear intersection and Bayesian change-point detection methods produce much wider ranges of warming onset estimates than those produced using the SiZer method on the same synthetic ensembles (Extended Data Fig. 3a–c). The linear methods are also less able to detect change-points for trends that are not simple linear functions. On the basis of our change-point method testing, we use the SiZer method in this study, because it appears to be most adaptable and stable for dealing with the climate changes that characterize industrial-era warming. We apply our method testing to assess the range of uncertainty in estimates of the onset of regional warming related to the SiZer change-point detection method (Extended Data Fig. 1b). This is carried out using an accelerating warming trend upon which AR(1) noise is added that has the same lag-1 autocorrelation and trend-to-variability characteristics as the regional reconstructions. These parameters are estimated by calculating characteristics of residuals about the 15-yr filters (as in Fig. 2a) of the regional reconstructions. Uncertainty in onset estimates related to the SiZer method is small, typically better than ±25 yr (5%–95%). Exceptions are Antarctica, for which the small trend relative to variability does not allow for the detection of significant trends, and Asia, for which strong lag-1 autocorrelation leads to uncertainty in the lower (5%) bound for the onset of warming (Extended Data Fig. 1b). We use the regional temperature reconstructions to determine the extent to which industrial-era warming has caused regional climates to emerge above the level of pre-industrial variability. We choose the interval ad 1622–1799 as the climate baseline to test emergence against. The starting point (ad 1622) represents the earliest year for which temperature reconstructions are available for all of the terrestrial and marine regions examined in this study. The end year (ad 1799) was chosen as a time well before the onset of industrial-era warming in any of the regional reconstructions (Table 1). It is also before strong volcanic cooling events associated with the 1809 Unknown and 1815 Tambora eruptions, which could skew the reference period towards cooler states. Time of emergence is detected when a climate change signal emerges above a defined noise threshold28. Here we use a threshold of 2σ of interannual variability above the mean of the reference interval (Fig. 2c). We smooth the reconstructions using filters with widths of 15–50 yr (the same as for our change-point assessments) and calculate the median time when the climate signal (smoothed reconstructions) emerges and stays above the noise threshold. The emergence year is quite insensitive to the level of smoothing we apply to the climate signal across the 15–50-yr filter widths, with the 5%–95% range of emergence estimates being between only 1 yr and 8 yr for the regional reconstructions for which climate emergence is found to occur. A difference between our climate emergence assessment and previously published time-of-emergence studies is that the palaeoclimate reconstructions allow us to assess climate emergence using a long baseline interval that occurs entirely before the onset of industrial-era warming. Our results demonstrate that, in some regions, industrial-era warming has already caused climate to emerge above the range of natural variability in the ad 1622–1799 reference interval (Table 1). We test the sensitivity of this result to the choice of reference interval and find that all tested reference intervals before the onset of industrial-era warming result in similar regional emergence patterns and timings, whereas reference periods that include parts of the industrial-era warming signal result in later estimates for the time of emergence. For example, in the Arctic reconstruction, time-of-emergence estimates based on different reference periods are: ad 1947 (ad 1500–1799 reference), ad 1930 (ad 1600–1799 reference), ad 1930 (ad 1622–1799 reference; Table 1), ad 1938 (ad 1700–1799 reference), ad 1960 (ad 1800–1899 reference) and ad 1978 (ad 1850–1899 reference). The overlapping interval of the regional reconstructions that we use for our time-of-emergence reference (ad 1622–1799) occurs during the time of coolest conditions during the past 2,000 years4, 15. Using longer palaeoclimate reference intervals that incorporate earlier, warm intervals of the past 2,000 years will alter time-of-emergence results, but it is not currently possible to perform this assessment consistently between regions, owing to the length limitations of currently available tropical ocean temperature reconstructions. We plot temperature trends from gridded instrumental datasets in Fig. 1. The surface air temperature datasets are from the Climate Research Unit (CRU) TS3.22 product56, or the ERAI-f product57 for surface air temperature over Antarctica, and the sea surface temperature datasets are from the HadiSST product58. The CO -equivalent record shown in Fig. 3d is derived from ref. 59. Matlab code for assessing the onset (and sign) of industrial-era climate trends using the SiZer method is archived with the NCEI World Data Service for Paleoclimatology at http://www.ncdc.noaa.gov/paleo/study/20083. These files include the original SiZer package24 obtained from http://www.unc.edu/~marron/marron_software.html, and additional code to assess the onset of sustained, significant temperature trends in annually resolved climate records.


Last year shattered 2014’s record to become the hottest year since reliable record-keeping began, two U.S. government science agencies announced Wednesday in yet another sign that the planet is heating up. 2015’s sharp spike in temperatures was aided by a strong El Niño weather pattern late in the year that caused ocean waters in the central Pacific to heat up. But the unusual warming started early and steadily gained strength in a year in which 10 of 12 months set records, scientists said. The new figures, based on separate sets of records kept by NASA and the National Oceanic and Atmospheric Administration, could fuel debate over climate change in an election year in which the two main political parties remain divided over what to do about global warming and, indeed, whether it exists. “2015 was by far the record year in all of the temperature datasets that are based on the instrumental and surface data,” said Gavin Schmidt, director of the Goddard Institute for Space Studies at NASA, which made the announcement jointly with NOAA, the National Oceanic and Atmospheric Administration. “It really underlines the fact that the planet really is still warming, there is no change in the long term global warming rate, and we know why that is,” he said. NASA reported that 2015 was officially 0.23 degrees Fahrenheit (0.13 degrees Celsius) hotter than 2014, the prior record year, a sharp increase for a global temperature record in which annual variation is often considerably smaller. NOAA’s figures showed slightly greater warming, of about 0.29 degrees Fahrenheit (0.16 degrees C) hotter than 2014. “A lot of times, you actually look at these numbers, when you break a record, you break it by a few hundredths of a degree,” said Thomas Karl, director of NOAA’s National Centers for Environmental Information. “But this record, we literally smashed. It was over a quarter of a degree Fahrenheit, and that’s a lot for the global temperature.” Overall, NOAA said, 2015 was 1.62 degrees Fahrenheit (0.9 degrees Celsius) above the 20th century average. NASA and NOAA both keep independent global surface temperature datasets, measuring temperatures over both the land and the oceans using thermometers, ocean buoys and ship readings. The datasets do not always agree perfectly, but they showed relatively little disagreement this year, Schmidt said. The latest record means that 2014 — the previous record year — only officially held that title for one year. 2014 came by its record by a relatively narrow margin — for instance, NASA gave 2014 a 38 percent chance of having been the warmest year on record, still reserving a nontrivial chance that the real warmest year had been 2010 or 2005. (NOAA gave a 48 percent chance that 2014 had, at the time, been the warmest year.) [Sorry, skeptics: NASA and NOAA were right about the 2014 temperature record] This year, in contrast, there is little need for citing percentages or a statistical photo finish. Buoyed by a powerful El Niño event, 2015 shattered the 2014 record. NASA’s Schmidt suggests there is only a 5 percent possibility that any other year on record was actually warmer. Fifteen of the 16 hottest years on record have now occurred in this century, according to NASA. U.S. officials stressed that the El Niño pattern alone does not account of the year’s record warmth. “The interesting thing is that 2015 did not start with an El Niño,” Schmidt said. “It was warm right from the beginning.” Because a strong El Niño still is in place, “2016 is expected to be an exceptionally warm year, and perhaps even another record,” Schmidt said. The release of the 2015 temperature data prompted statements from leading Democratic presidential candidates Hillary Clinton and Bernie Sanders. Clinton, in a Twitter posting, said, “Climate change is real. It’s hurting our planet and our people. We can’t afford a president who ignores the science.” The Sanders campaign also tweeted a response, saying. “Climate change is real and caused by human activity. This planet and its people are in trouble.” There was no immediate comments from the major GOP contenders, several of whom have been openly skeptical of the mainstream scientific view that human activity is causing the planet to warm. Front-runner Donald Trump has dismissed climate change as a hoax. According to the NOAA analaysis on Wednesday, every month in 2015 broke previous temperature records except for two: January and April. NOAA also announced Wednesday that for December, the “temperature departure from average was also the highest departure among all months in the historical record and the first time a monthly departure has reached 2°F.” From a climate policy perspective, the warmth of 2015 is also highly significant. Global leaders in Paris agreed in December that the planet should not be allowed to warm 2 degrees Celsius above pre-industrial temperatures — and ideally, warming should be limited to 1.5 degrees Celsius if possible. Based on 2015’s temperature record, though, we’re already half way to 2 degrees. “This is the first year where the record is clearly above 1 degree Celsius above the 19th century,” said NASA’s Schmidt. NOAA’s data also show that the planet is now more than 1 degree Celsius warmer than the average temperature between 1880 and 1899, said the agency’s Karl. 2015’s El Niño enhanced heat was accompanied by dramatic weather events across the globe, including a record for the number of Category 3 or greater tropical cyclones in the Northern Hemisphere. That tally includes Hurricane Patricia, the most intense hurricane ever recorded by the National Hurricane Center. [The Northern Hemisphere’s record shattering tropical cyclone season by the numbers] In some ways most ominously of all, 2015 was the year that scientists announced that an entirely new sector of Greenland — one containing over three feet of potential sea level rise — appeared to have been destabilized. The region is centered on the Zachariae and Nioghalvfjerdsfjorden glaciers of northeast Greenland, which together comprise the endpoint of the northeast Greenland ice stream, which drains 12 percent of the vast ice sheet. 2015’s record warmth also included a major anomaly — very cold temperatures in the North Atlantic Ocean to the south of Greenland. Monthly NOAA temperature maps repeatedly showed a blue colored “blob” of cold in this region, a development that is sparking increasing scientific interest, because of the suspicion that it could represent a sign of a change in the overturning circulation of the ocean. “In the northern North Atlantic, temperatures were colder than normal, and that was really pretty much the only part of the world that had a sizeable area with below average temperatures,” Karl said. [Why some scientists are worried about a surprisingly cold ‘blob’ in the North Atlantic Ocean] It certainly isn’t the case that the 2015 temperature record can be entirely attributed to the warming of the globe by human greenhouse gas emissions. Climate change has never meant that every successive year will be warmer than the last, and the powerful 2015 El Niño unlocked immense heat from the Pacific Ocean that drove up the global temperature. But at the same time, 2015 was also considerably hotter than 1998, another major El Niño year that was, at the time, the hottest year on record. Now, in contrast, it’s fifth or sixth on the list, depending on which agency you consult. And that, say experts, is how the warming of the planet makes itself felt. “It’s breaking the record because we also have this unusually strong El Niño, but at the same time we know the ocean is now absorbing two times more heat than around the last time we had a big El Niño, which is quite a while ago,” said Katharine Hayhoe, a climate scientist at Texas Tech University. There has been some talk in scientific circles that 2016 could be even hotter overall than 2015 — which would lead to three record years in a row. The reasoning here is that there is usually a significant lag between when El Niño peaks and when the warming of the globe does in its wake. Thus, 1998 was the hottest year of the 1997-1998 El Niño event. Britain’s Met Office recently forecast that 2016 could be “at least as warm, if not warmer” than 2015, in the words of research fellow Chris Folland. “In previous El Niño years, they peak in the wintertime … [and] the warmest temperatures are in the subsequent year,” said NOAA’s Karl.  “If 2016 continues like we’ve seen in the past, that would suggest 2016 is going to be very close to a record or even a new record.” However, not all scientists agree. “My guess is that 2016 may not be warmer than 2015,” said Kevin Trenberth, a climate change and El Niño expert at the National Center for Atmospheric Research. He thinks the current El Niño may already have begun to peak (or have peaked) and thus that the second half of 2016 may cool down again somewhat. In 2015, record warm temperatures and a growing focus on addressing global warming seemed in curious sync. It was the year that Pope Francis released his historic encyclical on the environment, Laudato Si, and the year in which the United States moved to regulate greenhouse gas emissions from the generation of electricity, their largest single source. Most significant, as heat records over the year accumulated, nations of the world assembled in Paris to forge a global climate agreement that will serve as the template for locking in cuts to greenhouse gas emissions in coming decades. It’s hard to say that 2015’s warmth directly contributed to these human decisions, and yet it’s also hard to entirely separate the two. The stark warming of the globe in 2015 clearly imparted a newfound sense of policy urgency. “NASA has been talking about the existence of global warming in public since 1988,” said Schmidt. “1988 was also a record warm year for the time. Just so that people understand, it is now 23rd in the rankings.” Scientists say human greenhouse gas emissions have canceled the next ice age Why clean energy is now expanding even when fossil fuels are cheap Why we’ve been hugely underestimating the overfishing of the oceans For more, you can sign up for our weekly newsletter here, and follow us on Twitter here.

Discover hidden collaborations