In 1941, New Mexico received almost twice its annual average statewide precipitation, setting two monthly precipitation records that still stand. What caused this anomaly?
The year 1941 stands out as by far the highest annual precipitation value in the climatological record in New Mexico (Fig. 1). The statewide annual precipitation for January–December 1941 was 67.5 cm, compared to a twentieth-century mean annual value of 35.4 cm. At first glance the value for 1941 looks like it might be an error in the data, which is not the case. Although annual precipitation values are not normally distributed—because extreme outliers like the 1941 data point cause positive skewness—the 1941 value is 4.8 standard deviations above the climatological annual mean. This study examines how, and to the extent possible from empirical evidence why, the year 1941 was so abnormally wet.
Annual precipitation time series for the state of New Mexico derived from climate divisional data (Vose et al. 2014). The datum for calendar year 1941 is 67.5 cm. The solid green line is the mean annual precipitation for the twentieth century (35.4 cm); dashed red lines delineate upper- and lower-quartile boundaries, such that half of the annual values for the twentieth century lie within the red lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Annual precipitation time series for the state of New Mexico derived from climate divisional data (Vose et al. 2014). The datum for calendar year 1941 is 67.5 cm. The solid green line is the mean annual precipitation for the twentieth century (35.4 cm); dashed red lines delineate upper- and lower-quartile boundaries, such that half of the annual values for the twentieth century lie within the red lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Annual precipitation time series for the state of New Mexico derived from climate divisional data (Vose et al. 2014). The datum for calendar year 1941 is 67.5 cm. The solid green line is the mean annual precipitation for the twentieth century (35.4 cm); dashed red lines delineate upper- and lower-quartile boundaries, such that half of the annual values for the twentieth century lie within the red lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Diagnosis of extreme weather and climate events is in the news frequently. Evidence for significant global temperature change over the past century is unequivocal (IPCC 2013), motivating intense scrutiny of possible signatures of long-term climate change in shorter-term extreme weather and climate events (e.g., IPCC 2012). For assessment of a climate anomaly on an annual scale, a central question is the extent to which an extreme anomaly (of total precipitation, in this case) was the sum of individual, unconnected weather events—analogous to an unusual, but random and unpredictable run of independent (presumably fair) dice throws at a casino. Alternatively, a series of weather events might be associated with one or more specific low-frequency causes—climate forcings—that acted to organize precipitation events over time. This distinction comes up frequently in the present day, associated with the debate over whether observed extreme anomalies might be attributed to long-term climate change (IPCC 2012).
Even a huge time-averaged anomaly may not require climatic forcing. A recent assessment of the severe 2012 “flash drought” on the U.S. Great Plains concluded that the drought did not require external forcings and could possibly have arisen from atmospheric noise alone, although the observed precipitation deficit was consistent with oceanic El Niño–Southern Oscillation (ENSO) forcing and antecedent dry soil moisture (Kumar et al. 2013). Hoerling et al. (2013) analyzed the extreme persistent drought/heat wave conditions centered on Texas in 2011, and similarly related the event in part to La Niña–related oceanic forcing. Land surface feedbacks seemed to play a significant role in these events, and strong covariance between temperature and precipitation (warm and dry) helps to explain the persistence of drought conditions.
The present study explores the opposite tail in the distribution of annual-scale precipitation anomalies, associated with extreme wetness. If drought consists of a persistent, often nearly continuous spell of no precipitation and extreme heat, then in contrast time-averaged pluvial anomalies may comprise a series of individual storms whose climatic connection may be less readily attributable. In arid regions just one or two very extreme storms can by themselves produce enough precipitation to generate above-average precipitation for an entire year, in which case the process of taking an annual total may be irrelevant to discussion of the cause of the anomaly.
Climate anomalies could be attributable, at least to some extent, if we understand the forcing and we understand the response to forcing in terms of weather events. One signal of a potentially forced anomaly is persistence over time scales longer than individual weather events. We are more likely to interpret an annual precipitation anomaly as “forced” if the anomaly consists of many individual events, instead of just one or a few very extreme storms. The annual time scale of the 1941 precipitation anomaly poses interesting challenges for climatic attribution studies, because prominent oceanic forcings (especially ENSO) and the atmospheric response to ENSO forcing seem strongly tied to specific seasons (e.g., Rasmusson and Carpenter 1982; Ropelewski and Halpert 1986; and many subsequent studies).
There could be multiple hypothesized causes of the 1941 precipitation anomaly that are not mutually exclusive. One of us (SMS) canvassed forecasters from the U.S. National Weather Service Office in Albuquerque about possible explanations for the well-known (though little studied) 1941 New Mexico precipitation anomaly before we started our analysis. We received multiple hypotheses, including the known presence of a strong El Niño event that year and the possibility that tropical cyclone remnants could have affected the state. Unlike recent studies of heat waves associated with drought cited above [see also Dole et al. (2014)], we will not focus much on global warming; there is no significant long-term trend in precipitation apparent in Fig. 1, and the most pronounced signal in temperature (globally and regionally) takes place in the late twentieth century (IPCC 2013), long after the 1941 anomaly of interest here.
We are also interested in examining local spatial variability within a time-averaged hydroclimatic anomaly. Even widespread precipitation events can be spatially complex, so that individual rain gauges record different events. Over a year’s time, a spatially averaged annual precipitation anomaly can be the result of quite different time series of daily precipitation events at observing sites within the region of the annual anomaly.
One of the challenges in diagnosing a regional climatic anomaly that occurred more than 70 years ago is that data sources were different, and generally much sparser, than data describing contemporary anomalies. In 1941 there were no satellite data, no operational radiosonde network, and sparse ocean surface observations. Hourly observations were only available for a single station in New Mexico (Albuquerque) starting on 1 July 1941.
Nevertheless, we can take advantage of surface data records and published indices of oceanic anomalies such as ENSO. The National Oceanic and Atmospheric Administration (NOAA) Central Library has a collection of U.S. daily surface weather maps for the period 1871–2004. The format of these maps changed in mid-1941 (on 1 August). Prior to that date, station plots are limited to wind direction, sky conditions, and weather; subsequently, they include air and dewpoint temperatures, changes in barometric pressure, wind velocity and direction, precipitation amount, synoptic features, and forecast discussions for cities across the continental United States. Historic references and data are available in the monthly Climatological Data publications from the U.S. Department of Commerce, Weather Bureau, as well as various surface charts and narrative summaries published in Monthly Weather Review. In addition, we utilize the Twentieth Century Reanalysis (20CR; Compo et al. 2006), a model-generated retrospective estimate of atmospheric circulation based on assimilated surface pressure data and observed SST anomalies.
QUANTIFICATION OF THE 1941 PLUVIAL YEAR.
To describe the spatial and temporal extent of the 1941 annual precipitation anomaly, we generated percentile maps of running averages of monthly Palmer drought severity index (PDSI) values calculated over the nine multistate U.S. climate regions. We defined annual-scale drought and pluvial regions as those with a 12-month average of PDSI for 1941 in the lower or upper tercile of all such annual averages (Fig. 2). The regions encompassing the West, west north-central, and Southwest United States experienced pluvial conditions in 1941 by this definition, while the central and Northeast regions were in drought. This large-scale pattern of wetness and dryness is very similar to the pattern associated with El Niño conditions in the tropical Pacific Ocean (Ropelewski and Halpert 1986).
Regional-scale pluvial (green), medial (gray), and drought (brown) conditions for calendar year 1941, defined in terms of historical terciles of 12-month average of monthly PDSI values. Regional boundaries are shown as heavy state boundary lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Regional-scale pluvial (green), medial (gray), and drought (brown) conditions for calendar year 1941, defined in terms of historical terciles of 12-month average of monthly PDSI values. Regional boundaries are shown as heavy state boundary lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Regional-scale pluvial (green), medial (gray), and drought (brown) conditions for calendar year 1941, defined in terms of historical terciles of 12-month average of monthly PDSI values. Regional boundaries are shown as heavy state boundary lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
A more detailed examination of the southwestern states exhibiting pluvial conditions in 1941 (Fig. 3) shows that the percentage of normal precipitation for climate divisions (subregions of each state that have relatively uniform climate characteristics) ranged from 100% to 300%. Every climate division in this six-state region received above-average annual precipitation in 1941. A swath of divisions extending westward across New Mexico from north and west Texas exhibited the largest anomalies. Precipitation anomalies decreased to average or below average north of Utah and Colorado and in eastern Texas and Oklahoma and were spatially variable across Arizona. Thus very wet annual average conditions in 1941 extended across a large swath of the southern Rockies, with the most extreme anomalies in New Mexico (dark blue divisions in Fig. 3).
Annual precipitation total in 1941, expressed as a percentage of twentieth-century annual average, in climate divisions in states across the southwestern United States. Letters a–d on the plot show locations of the sites used to illustrate daily precipitation statistics in Table 1 and Fig. 5.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Annual precipitation total in 1941, expressed as a percentage of twentieth-century annual average, in climate divisions in states across the southwestern United States. Letters a–d on the plot show locations of the sites used to illustrate daily precipitation statistics in Table 1 and Fig. 5.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Annual precipitation total in 1941, expressed as a percentage of twentieth-century annual average, in climate divisions in states across the southwestern United States. Letters a–d on the plot show locations of the sites used to illustrate daily precipitation statistics in Table 1 and Fig. 5.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
The 1941 anomaly was spread extraordinarily widely across the seasonal cycle, although wet anomalies did not persist uniformly from month to month. Positive precipitation anomalies across the southwestern United States appeared early in calendar year 1941 and regressed back to near normal the following winter, late in 1941 (Fig. 4; see also Fig. 9 to follow). Thus, calendar year 1941 represents a reasonable time frame for describing the duration of this pluvial anomaly. Annual results would not be much different if the 1941 water year, extending from October 1940 through September 1941, were used instead.
Monthly time series of statewide precipitation (cm) in New Mexico in 1941 (light blue line), superimposed on the historical (1901–2000) distribution of monthly statewide precipitation shown as a box and whiskers for each month. The horizontal line and diamond within each box denote the twentieth-century median and mean, respectively. Boxes delineate 25th and 75th percentiles, and whiskers extend to record minimum and maximum values for each month.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Monthly time series of statewide precipitation (cm) in New Mexico in 1941 (light blue line), superimposed on the historical (1901–2000) distribution of monthly statewide precipitation shown as a box and whiskers for each month. The horizontal line and diamond within each box denote the twentieth-century median and mean, respectively. Boxes delineate 25th and 75th percentiles, and whiskers extend to record minimum and maximum values for each month.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Monthly time series of statewide precipitation (cm) in New Mexico in 1941 (light blue line), superimposed on the historical (1901–2000) distribution of monthly statewide precipitation shown as a box and whiskers for each month. The horizontal line and diamond within each box denote the twentieth-century median and mean, respectively. Boxes delineate 25th and 75th percentiles, and whiskers extend to record minimum and maximum values for each month.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
In the New Mexico statewide average, 8 out of 12 months in 1941 were in the upper quartile of the historical distribution of precipitation—that is, very anomalously wet (Fig. 4). The transition seasons, spring and autumn, were both exceptionally wet in 1941. Two months (May and September) represent the record monthly statewide precipitation values for the month in the twentieth-century record; March is second wettest. Three additional months in 1941 (January, April, and October) are in the top 10% of twentieth-century precipitation in the state. November was the only month in 1941 with below-average statewide precipitation in New Mexico.
As shown by the quartile boxes in Fig. 4, maximum precipitation in New Mexico occurs climatologically in July and August, in association with the North American monsoon (Douglas et al. 1993). In 1941, however, precipitation in the summer monsoon months was only slightly above average, representing an unusual seasonal minimum in precipitation between spring and fall.
Numerous impacts associated with extreme wet conditions in 1941 are described in the narrative sections of the Climatological Data monthly and annual summary publications (Weather Bureau 1941). Exceptional precipitation and soil moisture are referenced in the summaries for several months, most notably in May, September, and October. The year ended with 28 documented weather-related deaths in New Mexico—24 of them attributed to floods or flash floods. Reservoir storage data indicate that Elephant Butte Reservoir on the Rio Grande in southern New Mexico was filled to capacity for the first time, reaching a storage record that still stands. Wetter-than-normal conditions were initially expected to produce above-average crop yields, but by October it was noted that harvests were greatly hindered by the wet weather. The annual summary (Hardy 1941) states that “the unprecedented wetness... completely overshadowed any previous wet year in the half century of climatological records of this state,” and nearly three-quarters of a century later that statement remains true.
WEATHER EVENTS CONTRIBUTING TO THE 1941 ANNUAL ANOMALY.
Monthly or annual precipitation anomalies represent accumulations of individual events that often are short in duration and cover limited areas. A finer-scale distribution of precipitation in space and time is illustrated using daily time series from Albuquerque and Roswell (Fig. 5, top and bottom, respectively). These cities are separated by ∼300-km distance (locations shown in Fig. 3) and by two substantial mountain ranges. It is therefore not surprising that some precipitation events occur at one site and not the other, especially in summer when isolated thunderstorms can drop significant rainfall over very limited areas. However, several major multiday precipitation events in spring and autumn months are evident in both time series in Fig. 5, with some of the daily precipitation totals exceeding long-term monthly averages.
Daily and monthly precipitation (cm) observed at (top) Albuquerque and (bottom) Roswell NM for the period 1 Jan–31 Dec 1941. Locations of Albuquerque and Roswell are shown by letters “a” and “b,” respectively, in Fig. 3. Blue vertical bars show local daily precipitation amounts relative to the right-hand ordinate. Green shading shows local monthly accumulations relative to the left-hand ordinate; green dashed lines show the local climatological monthly mean. Black horizontal lines indicate statewide estimates of monthly precipitation in 1941 (also shown as the blue “x”s in Fig. 4), repeated here in both panels, relative to the left-hand ordinate. The number of days with measurable precipitation each month at these sites is listed in Table 1.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Daily and monthly precipitation (cm) observed at (top) Albuquerque and (bottom) Roswell NM for the period 1 Jan–31 Dec 1941. Locations of Albuquerque and Roswell are shown by letters “a” and “b,” respectively, in Fig. 3. Blue vertical bars show local daily precipitation amounts relative to the right-hand ordinate. Green shading shows local monthly accumulations relative to the left-hand ordinate; green dashed lines show the local climatological monthly mean. Black horizontal lines indicate statewide estimates of monthly precipitation in 1941 (also shown as the blue “x”s in Fig. 4), repeated here in both panels, relative to the left-hand ordinate. The number of days with measurable precipitation each month at these sites is listed in Table 1.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Daily and monthly precipitation (cm) observed at (top) Albuquerque and (bottom) Roswell NM for the period 1 Jan–31 Dec 1941. Locations of Albuquerque and Roswell are shown by letters “a” and “b,” respectively, in Fig. 3. Blue vertical bars show local daily precipitation amounts relative to the right-hand ordinate. Green shading shows local monthly accumulations relative to the left-hand ordinate; green dashed lines show the local climatological monthly mean. Black horizontal lines indicate statewide estimates of monthly precipitation in 1941 (also shown as the blue “x”s in Fig. 4), repeated here in both panels, relative to the left-hand ordinate. The number of days with measurable precipitation each month at these sites is listed in Table 1.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
The number of days of measurable precipitation confirms that precipitation was unusually persistent and widespread from January through October 1941 (Table 1, which includes data from two additional sites). All four of these sites registered record high annual numbers of days with precipitation in 1941. Between January and May, multiple months at each site, and multiple sites in each month, exhibited at least double the number of days of precipitation expected from the long-term monthly mean. After merely slightly more frequent rains than normal in the summer months, September and October again featured more than twice the normal frequency of precipitation at multiple sites across the state.
Number of days of measurable precipitation (greater than a trace) for each month in 1941 at four sites in New Mexico, compared to the mean number of days of precipitation at each site. Values in 1941 that exceeded the long-term monthly mean are starred, values in 1941 that were double or more the long-term monthly mean are double starred, and record values are shown in boldface. Site locations are shown in Fig. 3. Daily precipitation time series for sites “a” and “b” (Albuquerque and Roswell, respectively) are shown in Fig. 5.
Visual examination of daily weather maps and 20CR upper-air reconstructions for March and April 1941 indicate that the late-winter storm track was exceptionally active. At the surface, persistent low pressure over the Southwest, combined with a strong surface high over the Great Plains, provided moist low-level easterly upslope flow across eastern New Mexico. Locally heavy amounts of rain and reports of hail suggest that convective processes were in place during this period. This synoptic situation is typical for cold season precipitation from frontal storms in New Mexico; in 1941 this regime was exceptionally persistent, consistent with other El Niño years (discussed further in the next section).
Later in the year, sea level pressure anomalies indicate low-level moisture from the Gulf of Mexico in May and a Pacific moisture surge in September likely contributed to anomalously high precipitation values. During the active periods 1–4 and 18–31 May, precipitation impacted most areas of New Mexico but favored the eastern plains, suggesting that both midlevel Pacific moisture and low-level return flow from the Gulf of Mexico contributed to the observed precipitation. Upper-level analyses produced by 20CR indicate active storm tracks during these two periods, with an upper-level trough or closed low west or over New Mexico.
Another exceptionally active storm period occurred during the latter half of September. The U.S. Daily Weather Map for 23 September indicates that rain fell across eastern New Mexico in a band extending southward over west Texas (Fig. 6, obtained from www.lib.noaa.gov/collections/imgdocmaps/daily_weather_maps.html). A cold front is analyzed southward from Albuquerque, with occluded and warm fronts crossing the state from southwest to northeast. (Occluded fronts are almost never shown in the southwestern United States in modern analyses.) A tropical storm is shown, centered just off the Gulf Coast of Texas, associated with rain along the coast that does not extend continuously across Texas. Airmass characteristics in central Texas are labeled as “MP + MT,” suggesting a combination of midlatitude and tropical moist air masses. Dewpoint temperatures exceed 60°F from southeastern New Mexico through central Texas with east–southeast surface flow supporting a steady supply of low-level moisture into southeastern New Mexico.
U.S. Weather Bureau surface weather map for 23 Sep 1941. The portion of the map covering the southern plains states (New Mexico, Texas, Oklahoma) is expanded across the lower right of the plot. Station observations represent conditions at 0130 EST. Analyses include mean sea level isobars, fronts and airmass classifications, 12-h hurricane tracks, and shading to represent areas of precipitation at the time of observation. Precipitation values for the previous 18 h are included in the station data. The red annotations were added manually after publication and prior to digitizing, marking stations that received precipitation.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
U.S. Weather Bureau surface weather map for 23 Sep 1941. The portion of the map covering the southern plains states (New Mexico, Texas, Oklahoma) is expanded across the lower right of the plot. Station observations represent conditions at 0130 EST. Analyses include mean sea level isobars, fronts and airmass classifications, 12-h hurricane tracks, and shading to represent areas of precipitation at the time of observation. Precipitation values for the previous 18 h are included in the station data. The red annotations were added manually after publication and prior to digitizing, marking stations that received precipitation.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
U.S. Weather Bureau surface weather map for 23 Sep 1941. The portion of the map covering the southern plains states (New Mexico, Texas, Oklahoma) is expanded across the lower right of the plot. Station observations represent conditions at 0130 EST. Analyses include mean sea level isobars, fronts and airmass classifications, 12-h hurricane tracks, and shading to represent areas of precipitation at the time of observation. Precipitation values for the previous 18 h are included in the station data. The red annotations were added manually after publication and prior to digitizing, marking stations that received precipitation.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
September is the most likely month for the southwestern United States to be directly affected by tropical cyclone remnants, most often originating in the northeastern Pacific (e.g., Ritchie et al. 2011). The year 1941 was an above-average month for eastern Pacific tropical cyclone activity, and September 1941 was especially active with three storms tracking northward along the west coast of Mexico (Lewis 2003).
The surface map for 23 September 1941 does not portray a signature of a cyclone remnant from the eastern Pacific, although Hurd (1941) reports evidence of a very strong cyclone that crossed the Gulf of California on 20 September, which could have affected moisture transport northward into the southwestern United States in subsequent days. Additionally, details in the standard monthly narrative (Monthly Weather Review, 1941, Vol. 69, No. 9) for “Weather on the North Pacific Ocean” describe conditions associated with two tropical cyclones west of Mexico, the second of which occurred during the period 17–20 September. A press release from Mexico City, Mexico, described this event as “the fiercest cyclone of the century” resulting in considerable loss of life and property on the Baja Peninsula. One ship reported extreme hurricane conditions and pressure falls, with the second lowest barometer reading recorded to that date in the eastern Pacific. While the narrative also states that no additional information was known regarding the subsequent history of the cyclone, the remnants of the system could very likely have impacted New Mexico.
We augmented the surface weather maps with 20CR upper-air maps (Fig. 7), keeping in mind that the upper-air analyses are not anchored by any actual measurements above the surface and that surface pressure observations in northern Mexico are very sparse. 20CR includes a deep low at 500 hPa over the western United States and a clear pattern of lower-tropospheric humidity and temperature gradients typical of a baroclinic frontal system associated with the 500-hPa trough. This pattern is similar to a recurving track pattern described by Ritchie et al. (2011), in which a North Pacific tropical cyclone remnant interacts with a southward-digging midlatitude trough. A comparison of September surface wind observations (hourly for Albuquerque and daily for Roswell) with wind statistics for the years 1985 through 2005 suggests anomalous southwest flow at both stations but especially Roswell.
20CR output for 23 Sep 1941 across North America, corresponding to the surface weather map in Fig. 6. (a) Geopotential height, 500 hPa (m). (b) Air temperature, 500 hPa (K). (c) Specific humidity, 700 hPa (kg kg–1). (d) Specific humidity, 2 m above surface (kg kg–1).
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
20CR output for 23 Sep 1941 across North America, corresponding to the surface weather map in Fig. 6. (a) Geopotential height, 500 hPa (m). (b) Air temperature, 500 hPa (K). (c) Specific humidity, 700 hPa (kg kg–1). (d) Specific humidity, 2 m above surface (kg kg–1).
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
20CR output for 23 Sep 1941 across North America, corresponding to the surface weather map in Fig. 6. (a) Geopotential height, 500 hPa (m). (b) Air temperature, 500 hPa (K). (c) Specific humidity, 700 hPa (kg kg–1). (d) Specific humidity, 2 m above surface (kg kg–1).
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
It appears that exceptionally heavy precipitation in eastern New Mexico at this time resulted from the interaction of one tropical cyclone remnant—or perhaps two—with a midlatitude system, generating record-setting rainfall across the eastern plains of New Mexico that extended northward along the eastern slopes of the southern Rocky Mountains. This combination of tropical and extratropical dynamics represents a “perfect storm” for very heavy autumn season precipitation in the southwestern United States.
OCEANIC FORCING AND LARGE-SCALE CIRCULATION ANOMALIES.
Individual synoptic-scale weather events can be organized by large-scale circulation anomalies and known atmospheric forcings. These circulation features and forcings potentially provide a framework for understanding anomalous time-averaged weather, such as exceptionally wet conditions that persist for weeks or months as in 1941. Across the southwestern United States, these features include the winter storm track, an average latitude of west-to-east baroclinic wave occurrence that is known to be affected by large-scale temperature anomalies in the Pacific Ocean (Higgins et al. 2000; Seager and Hoerling 2014) during the time of year when such waves are active in middle latitudes (mostly in winter and spring). In the summer season, the North American monsoon circulation (Adams and Comrie 1997; Higgins et al. 1997) organizes moist convection across the region.
Large-scale climate anomalies in the transition seasons have not been studied so intensively. In autumn, tropical cyclone remnants as discussed above produce copious precipitation as seemed to occur in late September 1941. Individual cyclones are not predictable on climatic time scales, but some research has attempted to determine whether the annual frequency of cyclone remnants affecting the Southwest is correlated with large-scale forcing (Larson et al. 2005).
Pacific Ocean temperature variations.
ENSO, described in terms of tropical Pacific SST and air pressure, is one of the most important coupled ocean–atmosphere phenomena associated with global climate variability on interannual time scales. Changes in Pacific SST result in modified atmospheric circulations that can change the positions of the subtropical jet stream and storm tracks and, therefore, the distribution and magnitude of precipitation across the United States. The southern United States, including New Mexico, receives more precipitation than usual during El Niño, particularly during the cool seasons of winter and spring (Ropelewski and Halpert 1986; Redmond and Koch 1991).
The year 1941 was one of stronger El Niño events in the twentieth-century record in terms of the Niño-3.4 SST index, marking the peak of a multiyear warm phase that extended between 1939 and 1942 (Fig. 8, top). The North Atlantic Oscillation (NAO; Hurrell 1995) was also exceptionally perturbed in 1941, exhibiting one of the most negative index values of the twentieth century that year. Brönnimann et al. (2004) attributed the particularly severe winter of 1941 in Europe to the combination of intense El Niño and negative NAO conditions, mediated by the effects of these oceanic anomalies on the stratospheric circulation.
(top) Time series of monthly Niño-3.4 (gray line) and PDO (green fill) indices for the period 1901–2000. Niño-3.4 values greater than 1.0 (less than –1.0) are shaded red (blue). Yellow shading denotes the period from Nov 1940 to Feb 1942, for comparison to Fig. 9. (middle) SST annual anomaly from long-term mean for calendar year 1941. Contour interval 0.5°C; yellow–red colors indicate positive anomalies and green–blue–purple colors indicate negative anomalies (data obtained from NOAA/Earth System Research Laboratory). (bottom) Scatterplot of annual values of Mar–Sep averages of Niño-3.4 and PDO indices, with contemporaneous New Mexico precipitation values indicated by the size of each dot. The 1941 data point is the large red dot.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
(top) Time series of monthly Niño-3.4 (gray line) and PDO (green fill) indices for the period 1901–2000. Niño-3.4 values greater than 1.0 (less than –1.0) are shaded red (blue). Yellow shading denotes the period from Nov 1940 to Feb 1942, for comparison to Fig. 9. (middle) SST annual anomaly from long-term mean for calendar year 1941. Contour interval 0.5°C; yellow–red colors indicate positive anomalies and green–blue–purple colors indicate negative anomalies (data obtained from NOAA/Earth System Research Laboratory). (bottom) Scatterplot of annual values of Mar–Sep averages of Niño-3.4 and PDO indices, with contemporaneous New Mexico precipitation values indicated by the size of each dot. The 1941 data point is the large red dot.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
(top) Time series of monthly Niño-3.4 (gray line) and PDO (green fill) indices for the period 1901–2000. Niño-3.4 values greater than 1.0 (less than –1.0) are shaded red (blue). Yellow shading denotes the period from Nov 1940 to Feb 1942, for comparison to Fig. 9. (middle) SST annual anomaly from long-term mean for calendar year 1941. Contour interval 0.5°C; yellow–red colors indicate positive anomalies and green–blue–purple colors indicate negative anomalies (data obtained from NOAA/Earth System Research Laboratory). (bottom) Scatterplot of annual values of Mar–Sep averages of Niño-3.4 and PDO indices, with contemporaneous New Mexico precipitation values indicated by the size of each dot. The 1941 data point is the large red dot.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Furthermore, the Pacific decadal oscillation (PDO) or PDO index (Mantua et al. 1997) reached its maximum value in the historical record in 1941 (Fig. 8, top). The pattern of SST anomalies in 1941 (Fig. 8, middle) confirms that both ENSO and PDO were in high-magnitude realizations of their warm phases, including very anomalously cold temperatures in the North Pacific. Statistically, the climate signal of El Niño on North American precipitation is likely to be stronger when the PDO is in its positive phase (Gershunov and Barnett 1998; McCabe and Dettinger 1999; Higgins et al. 2000; Gutzler et al. 2002). The dynamical basis for ENSO–PDO relationships is a topic of active research (e.g., Newman et al. 2003; Sun and Bryan 2010), but it seems notable here that extremely persistent multimonth Pacific storm-track anomalies coincided with a record high PDO index in 1941.
North American monsoon and warm season anomalies.
Almost 50% of Albuquerque’s annual climatological precipitation is received during the summer monsoon season (Douglas et al. 1993; Fig. 4). Based on a ranking of statistics of month-by-month monsoon rainfall in Albuquerque for the period 1892–2013 compiled by the U.S. National Weather Service Tucson, the 1941 monsoon was ranked the 20th highest total rainfall for June–September precipitation with a total of 5.44 in. (13.82 cm), most of which fell in July (5.46 cm) followed by 4.70 cm in September 1941. These are not unusual monthly anomaly values (Fig. 4).
Monsoon precipitation usually has a well-defined onset that in most years marks the end of the pronounced spring dry season in the Southwest (Higgins et al. 1997). However, this was not the case in 1941, given the record-setting precipitation in May. Using the National Weather Service’s semiofficial criterion of three consecutive days of dewpoints above 47°F in New Mexico (Ellis et al. 2004), the average onset date is 3 July in southwestern New Mexico and 9 July for Albuquerque. Onset in 1941 by this definition was approximately 22 June, an early onset. Early onset dates generally lead to more total precipitation during the monsoon season (Higgins et al. 1998), as happened in 1941. However, the core monsoon rainfall months of July and August were bookended in 1941 by extreme anomalies in May and September, so that the summer months actually represented an extraordinarily unusual seasonal minimum in precipitation in 1941.
The role of ENSO in forcing North American monsoon anomalies is an ongoing source of debate. Several analyses based on late twentieth-century data have suggested that positive summer precipitation anomalies follow deficient snowpack (Gutzler 2000) or springtime La Niña conditions (Castro et al. 2001); neither of these concepts is consistent with the 1941 anomaly. Other studies suggest the opposite relationship, with El Niño associated with abundant monsoonal precipitation (Ropelewski and Halpert 1986; Hereford and Webb 1992).
As noted previously, a tropical cyclone played a role in heavy precipitation in late September in New Mexico. Tropical cyclone remnants represent a largely unpredictable source of late-summer precipitation that surely added to the overall wet conditions across the Southwest in 1941. Larson et al. (2005) noted an association between late-summer anomalies in the Arctic Oscillation (Thompson and Wallace 2000) and landfalling North Atlantic tropical cyclones but the relevance of this relationship to 1941 cyclone activity is difficult to quantify.
Was the 1941 precipitation anomaly forced?
The Northern Hemisphere oceans were in a highly anomalous state in 1941. Pacific SST anomalies present a strong statistical expectation that 1941 should be a high-precipitation year across the southwestern United States in the winter and spring. The roles of ocean temperature anomalies outside the tropical Pacific, or in summer and autumn seasons, on Southwest precipitation are ambiguous, however. The NAO shows no statistically significant signal with western North American precipitation. The PDO index is strongly correlated with cold season precipitation on decadal time scales (Gershunov and Barnett 1998; Gutzler et al. 2002). Whether this is due to the extratropical component of PDO, or whether PDO simply represents “decadal ENSO,” is a topic of active debate (Newman et al. 2003; Sun and Bryan 2010; Seager and Hoerling 2014).
However, the magnitude of the 1941 precipitation anomaly was extraordinary relative to these Pacific SST forcings compared to other ENSO years in the twentieth century (Fig. 8, bottom). There is a significant, but not very strong, linear correlation between New Mexico annual statewide precipitation and either PDO (r = 0.47) or Niño-3.4 (r = 0.43), but the three most extreme years of joint positive PDO and Niño-3.4 indices are associated with heavy precipitation.
A comparison of monthly precipitation between 1941 and the five most comparable high-PDO/El Niño years in the twentieth century (Fig. 9) shows that the autumn (September–October) precipitation anomalies in 1941 (blue line) lie far outside the envelope of precipitation in the comparable years (red dashed lines). The seasonal timing of the largest monthly anomalies in late spring and autumn does not conform to the peak in covariance between Pacific SST and continental precipitation in late fall–early spring (Redmond and Koch 1991). It is difficult to attribute the transition season precipitation anomalies in 1941 directly to SST forcing based on historical analogs. The lack of a strong and consistent ENSO/PDO signal in New Mexico precipitation in summer and autumn months is evident in Fig. 9.
Monthly precipitation averaged over the state of New Mexico for six years with high Niño-3.4 and PDO indices (1941 in solid blue; 1905, 1983, 1987, 1993, and 1997 shown as dashed red lines). Precipitation values for the 12 months in these years, plus the preceding Nov–Dec and subsequent Jan–Feb months, are plotted. The climatological 33% and 67% exceedance values (the lower-tercile and upper-tercile thresholds, respectively, based on the twentieth-century record) for each month are shown as thick black lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Monthly precipitation averaged over the state of New Mexico for six years with high Niño-3.4 and PDO indices (1941 in solid blue; 1905, 1983, 1987, 1993, and 1997 shown as dashed red lines). Precipitation values for the 12 months in these years, plus the preceding Nov–Dec and subsequent Jan–Feb months, are plotted. The climatological 33% and 67% exceedance values (the lower-tercile and upper-tercile thresholds, respectively, based on the twentieth-century record) for each month are shown as thick black lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
Monthly precipitation averaged over the state of New Mexico for six years with high Niño-3.4 and PDO indices (1941 in solid blue; 1905, 1983, 1987, 1993, and 1997 shown as dashed red lines). Precipitation values for the 12 months in these years, plus the preceding Nov–Dec and subsequent Jan–Feb months, are plotted. The climatological 33% and 67% exceedance values (the lower-tercile and upper-tercile thresholds, respectively, based on the twentieth-century record) for each month are shown as thick black lines.
Citation: Bulletin of the American Meteorological Society 97, 6; 10.1175/BAMS-D-14-00142.1
We examined the ensemble of Twentieth Century Reanalysis simulations to see if the global ocean anomalies during 1941, which are prescribed for the 20CR simulations, reproduce anything close to the exceptional precipitation observed across southwest North America, especially during the transition seasons. The ensemble mean precipitation for March through May 1941 is anomalously wet, consistent with expectations for El Niño conditions in the tropical Pacific. But the simulated precipitation anomalies in 1941 are well within the envelope of variability associated with other El Niño spring seasons: March–May is considerably wetter in 1992 compared to 1941 in the 20CR ensemble, for example. Simulated ensemble mean precipitation in autumn months in 1941 was not anomalous at all, with considerable spread among ensemble members. These results suggest that unforced variability plays a significant role in 1941 precipitation. The 23 September event discussed in the “Weather events contributing to the 1941 annual anomaly” section illustrates that the surface signature of at least one important weather feature—the Pacific hurricane remnant documented in Monthly Weather Review (1941)—was not present in the boundary values used to anchor the simulations.
So, while ENSO and possibly other large-scale patterns of oceanic variability almost certainly played a prominent role in the 1941 continental precipitation anomaly, observational and dynamical attribution metrics are insufficient to explain the record-setting magnitude and persistence of the regional precipitation anomaly. A considerable fraction of precipitation events, especially in the transition seasons, is not clearly attributable to oceanic forcing.
DISCUSSION AND CONCLUSIONS.
The 1941 annual precipitation anomaly extended across a broad swath of the southwestern United States, with the largest anomalies relative to climatological averages in New Mexico. The annual anomaly was notable for its persistence across the seasonal cycle extending into the autumn months. The largest monthly precipitation anomalies in New Mexico occurred in transition season months that do not fit easily into current physical paradigms of seasonal anomalies (e.g., winter storm-track variations or summer monsoon variability). This feature of the annual anomaly makes it difficult to attribute to documented modes of atmospheric forcing.
With regard to forcing, the extremely pronounced positive anomalies in ENSO and PDO indices, including values in late spring and summer, are qualitatively consistent with above-average precipitation across the Southwest in other years. Multiple storm events in the late spring are consistent with a southward displaced storm track and favorably positioned troughs. Nevertheless, the magnitude of anomalous precipitation statewide still exceeded, by far, the precipitation observed in the most comparable El Niño years in the historical record. The 1941 El Niño persisted unusually late into the spring and early summer, making it difficult to use statistics from post–World War II El Niño events as analogs for precipitation anomalies in 1941. The Niño-3.4 and PDO indices remained at historically positive values much later in the calendar year compared to many other El Niño years, roughly coincident with wet months in New Mexico. This year serves as a reminder that the relatively rich data record from the second half of the twentieth century is undoubtedly an incomplete sample of the natural variability of climate. Forced atmospheric model experiments with 1941-based oceanic boundary conditions could provide more definitive insights into the role of SST on the extreme annual precipitation anomaly documented here.
The availability of weather data for events in 1941 limits our ability to explore in detail the meteorological evolution of that very wet year. However, an increasingly rich variety of data sources is available (not all of which have been discussed in this paper), if not always obviously accessible. The huge and obvious gap in data associated with direct upper-air data cannot be filled, but dynamical proxies such as the Twentieth Century Reanalysis can play a very useful role in suggesting plausible three-dimensional circulation systems consistent with surface observations. We hope that studies such as this one help to motivate expanded and sustained archives of maps such as shown in Fig. 6 and other sources of historical data such as the SST data utilized for Figs. 8 and 9.
Our assessment of the role of forcing for this annual precipitation anomaly, derived principally from Twentieth Century Reanalysis simulations, should be considered as merely suggestive. The roles of oceanic variability associated with ENSO and PDO indices could be explored in more detail with controlled simulations. In particular, the roles of North Pacific and Atlantic Ocean temperature anomalies in shifting seasonal storm tracks, and the seasonality of far-field ENSO response in late-spring and autumn months, represent fruitful topics for future investigation.
So was the annual anomaly something forced, and therefore potentially predictable if the appropriate forcing could be predicted, or was it a random fluke? We cannot attribute the magnitude of the annual precipitation anomaly to identifiable climate forcing. But the existence of widespread precipitation anomalies, month after month, is strongly suggestive of forced variability. ENSO and PDO anomalies were both positive, qualitatively consistent with enhanced precipitation across the southwestern states as observed. The coexistence of positive ENSO and PDO reinforces the tendency for anomalously high precipitation, although the dynamical basis for considering PDO forcing distinct from ENSO remains ambiguous.
In 1941, the wettest year in the instrumented history of New Mexico coincided with one of the most pronounced El Niño–PDO anomalies in the twentieth century. With perfect hindsight, we might expect 1941 to have been a wet year. However, the extreme magnitude of the precipitation anomaly, the timing of precipitation anomalies in the transition season months, and the remarkable persistence of wet months across the seasonal cycle do not seem entirely consistent with ENSO forcing as described based on late twentieth-century variability. We might attribute some component of the wet anomaly to climatic forcing, but unattributable extreme weather contributed to the record precipitation as well.
ACKNOWLEDGMENTS
SMS acknowledges NWS Albuquerque for support as a student volunteer and the American Meteorological Society for an undergraduate scholarship. Dr. Ariane Pinson (USACE Albuquerque) provided helpful comments on an early version of the manuscript. Dr. Gil Compo provided helpful advice with the Twentieth Century Reanalysis. Support for the 20CR Project dataset is provided by the U.S. Department of Energy, Office of Science Innovative and Novel Computational Impact on Theory and Experiment program, and Office of Biological and Environmental Research, and by the NOAA Climate Program Office. Kaplan SST V2 data provided by NOAA/OAR/ESRL PSD, Boulder, Colorado (www.esrl.noaa.gov/psd/). Dr. Ken Kunkel, BAMS Editor Mike Alexander, and two anonymous reviewers provided exceptionally thorough and helpful suggestions for revision.
REFERENCES
Adams, D. K., and A. C. Comrie, 1997: The North American monsoon. Bull. Amer. Meteor. Soc., 78, 2197–2213, doi:10.1175/1520-0477(1997)078<2197:TNAM>2.0.CO;2.
Brönnimann, S., J. Luterbacher, J. Staehelin, T. M. Svendby, G. Hansen, and T. Svenøe, 2004: Extreme climate of the global troposphere and stratosphere in 1940–42 related to El Niño. Nature, 431, 971–974, doi:10.1038/nature02982.
Castro, C. L., T. B. McKee, and R. A. Pielke Jr., 2001: The relationship of the North American monsoon to tropical and North Pacific sea surface temperatures as revealed by observational analyses. J. Climate, 14, 4449–4473, doi:10.1175/1520-0442(2001)014<4449:TROTNA>2.0.CO;2.
Compo, G. P., J. S. Whitaker, and P. D. Sardeshmukh, 2006: Feasibility of a 10-year reanalysis using only surface pressure data. Bull. Amer. Meteor. Soc., 87, 175–190, doi:10.1175/BAMS-87-2-175.
Dole, R., and Coauthors, 2014: The making of an extreme event: Putting the pieces together. Bull. Amer. Meteor. Soc., 95, 427–440, doi:10.1175/BAMS-D-12-00069.1.
Douglas, M. W., R. A. Maddox, K. Howard, and S. Reyes, 1993: The Mexican monsoon. J. Climate, 6, 1665–1677, doi:10.1175/1520-0442(1993)006<1665:TMM>2.0.CO;2.
Ellis, A. W., E. M. Saffell, and T. W. Hawkins, 2004: A method for defining monsoon onset and demise in the southwestern USA. Int. J. Climatol., 24, 247–265, doi:10.1002/joc.996.
Gershunov, A., and T. P. Barnett, 1998: Interdecadal modulation of ENSO teleconnections. Bull. Amer. Meteor. Soc., 79, 2715–2725, doi:10.1175/1520-0477(1998)079<2715:IMOET>2.0.CO;2.
Gutzler, D. S., 2000: Covariability of spring snowpack and summer rainfall across the Southwest United States. J. Climate, 13, 4018–4027, doi:10.1175/1520-0442(2000)013<4018:COSSAS>2.0.CO;2.
Gutzler, D. S., D. M. Kann, and C. Thornbrugh, 2002: Modulation of ENSO-based long-lead outlooks of Southwest U.S. winter precipitation by the Pacific decadal oscillation. Wea. Forecasting, 17, 1163–1172, doi:10.1175/1520-0434(2002)017<1163:MOEBLL>2.0.CO;2.
Hardy, E. L., 1941: New Mexico section. Climatological data, Vol. XLV, No. 13, Weather Bureau, 104 pp.
Hereford, R., and R. H. Webb, 1992: Historic variation of warm-season rainfall, Southern Colorado Plateau, Southwestern U.S.A. Climatic Change, 22, 239–256, doi:10.1007/BF00143030.
Higgins, R. W., Y. Yao, and X.-L. Wang, 1997: Influence of the North American monsoon system on the U.S. summer precipitation regime. J. Climate, 10, 2600–2622, doi:10.1175/1520-0442(1997)010<2600:IOTNAM>2.0.CO;2.
Higgins, R. W., K. C. Mo, and Y. Yao, 1998: Interannual variability of the United States summer precipitation regime. J. Climate, 11, 2582–2606, doi:10.1175/1520-0442(1998)011<2582:IVOTUS>2.0.CO;2.
Higgins, R. W., A. Leetmaa, Y. Xue, and A. Barnston, 2000: Dominant factors influencing the seasonal predictability of U.S. precipitation and surface air temperature. J. Climate, 13, 3994–4017, doi:10.1175/1520-0442(2000)013<3994:DFITSP>2.0.CO;2.
Hoerling, M., and Coauthors, 2013: Anatomy of an extreme event. J. Climate, 26, 2811–2832, doi:10.1175/JCLI-D-12-00270.1.
Hurd, W. E., 1941: Weather on the North Pacific Ocean. Mon. Wea. Rev., 69, 274–275, doi:10.1175/1520-0493(1941)069<0274:WOTNPO>2.0.CO;2.
Hurrell, J. W., 1995: Decadal trends in the North Atlantic Oscillation: Regional temperatures and precipitation. Science, 269, 676–679, doi:10.1126/science.269.5224.676.
IPCC, 2012: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. C. B. Field et al., Eds., Cambridge University Press, 582 pp.
IPCC, 2013: Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1535 pp., doi:10.1017/CBO9781107415324.
Kumar, A., M. Chen, M. Hoerling, and J. Eischeid, 2013: Do extreme climate events require extreme forcings? Geophys. Res. Lett., 40, 3440–3445, doi:10.1002/grl.50657.
Larson, J., Y. Zhou, and R. W. Higgins, 2005: Characteristics of landfalling tropical cyclones in the United States and Mexico: Climatology and interannual variability. J. Climate, 18, 1247–1262, doi:10.1175/JCLI3317.1.
Lewis, M. D., 2003: A comprehensive revision of eastern North Pacific tropical cyclone tracks, 1921-1966. M.S. thesis, Dept. of Atmospheric Sciences, Creighton University, 81 pp.
Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78, 1069–1079, doi:10.1175/1520-0477(1997)078<1069:APICOW>2.0.CO;2.
McCabe, G. J., and M. D. Dettinger, 1999: Decadal variations in the strength of ENSO teleconnections with precipitation in the western United States. Int. J. Climatol., 19, 1399–1410, doi:10.1002/(SICI)1097-0088(19991115)19:13<1399::AID-JOC457>3.0.CO;2-A.
Newman, M., G. P. Compo, and M. Alexander, 2003: ENSO-forced variability of the Pacific decadal oscillation. J. Climate, 16, 3853–3857, doi:10.1175/1520-0442(2003)016<3853:EVOTPD>2.0.CO;2.
Rasmusson, E. M., and T. H. Carpenter, 1982: Variations in the tropical sea surface temperature and surface wind fields associated with the Southern Oscillation/El Niño. Mon. Wea. Rev., 110, 354–384, doi:10.1175/1520-0493(1982)110<0354:VITSST>2.0.CO;2.
Redmond, K. T., and R. W. Koch, 1991: Surface climate and stream flow variability in the western United States and their relationship to large-scale circulation indices. Water Resour. Res., 27, 2381–2399, doi:10.1029/91WR00690.
Ritchie, E. A., K. M. Wood, D. S. Gutzler, and S. R. White, 2011: The influence of eastern Pacific tropical cyclone remnants on the southwestern United States. Mon. Wea. Rev., 139, 192–210, doi:10.1175/2010MWR3389.1.
Ropelewski, C. F., and M. S. Halpert, 1986: North American precipitation and temperature patterns associated with the El Niño/Southern Oscillation (ENSO). Mon. Wea. Rev., 114, 2352–2362, doi:10.1175/1520-0493(1986)114<2352:NAPATP>2.0.CO;2.
Seager, R., and M. P. Hoerling, 2014: Atmosphere and ocean origins of North American drought. J. Climate, 27, 4581–4606, doi:10.1175/JCLI-D-13-00329.1.
Sun, D.-Z., and F. Bryan, 2010: Extratropical air-sea interaction, sea surface temperature variability, and the Pacific Decadal Oscillation. Climate Dynamics: Why Does Climate Vary?, Geophys. Monogr., Vol. 189, Amer. Geophys. Union, 123–148, doi:10.1029/2008GM000794.
Thompson, D. W. J., and J. M. Wallace, 2000: Annular modes in the extratropical circulation. Part I: Month-to-month variability. J. Climate, 13, 1000–1016, doi:10.1175/1520-0442(2000)013<1000:AMITEC>2.0.CO;2.
Vose, R. S., and Coauthors, 2014: Improved historical temperature and precipitation time series for U.S. climate divisions. J. Appl. Meteor. Climatol., 53, 1232–1251, doi:10.1175/JAMC-D-13-0248.1.
Weather Bureau, 1941: Climatological data. Vol. XLV, 104 pp.
