The boreal spring variability of the Intra-Americas low-level jet and its relation with precipitation and tornadoes in the eastern United States
The Intra-Americas Sea (IAS) low-level jet has been studied mainly for the summer and winter seasons. In contrast, spring conditions have been studied less. Here we analyze the boreal spring variability of the IAS low-level jet (IA-LLJ) and its relation with precipitation and tornadic activity in the region of the lower Mississippi, Tennessee and Ohio River basins (MORB). The main mode of variability of the spring IA-LLJ is obtained from a combined principal component analysis of zonal and meridional winds at 925-hPa. The first empirical orthogonal function of the IA-LLJ is a strengthening of the climatological flow with stronger easterlies in the Caribbean and stronger southeasterlies in the Gulf of Mexico. This first mode of variability of the IA-LLJ is related mainly to the Pacific North American (PNA) teleconnection pattern as the PNA modulates the pressure in the southeast region of the U.S. Consequently, there is an increase in precipitation over the MORB region as the moisture fluxes associated with the IA-LLJ increase. Tornadic activity in nine states spanning the MORB region is also significantly related to the IA-LLJ and the PNA index for March, in addition to the Pacific decadal oscillation (PDO) and the Niño indexes. Among the environmental factors that influence tornadic activity are southwesterly wind shear, dry transients at the mid-troposphere, moist transients at low levels, and an increase in convective available potential energy (CAPE). The decadal shifts in MORB precipitation and tornado activity appear to be related to the decadal shift of the IA-LLJ.
KeywordsLow-level jet Tornadoes PNA Mississippi River Intra-Americas Sea
The Intra-Americas Sea (IAS; Gulf of Mexico and Caribbean Sea) is a source of moisture for the eastern region of the United States of America (US). Previous studies by Rasmusson (1967), Bosilovich and Schubert (2002), Mestas-Nuñez et al. (2007) and Ruiz-Barradas and Nigam (2005) highlight the importance of the influence of IAS atmospheric moisture on the moisture budget of the continental US. The bulk of the moisture transport from the IAS is at the lower-levels below 700 hPa (Helfand and Schubert 1995). It travels through a low-level corridor comprised of the Caribbean easterlies and the Gulf of Mexico southeasterlies. In fact, it is the easterly Caribbean low-level jet (Muñoz et al. 2008) and the Gulf of Mexico southeasterlies (Mestas-Nuñez et al. 2007) that are the main wind components of the IAS moisture corridor.
The focus of the studies mentioned in the previous paragraph has been on the summer and winter seasons, whereas the spring season has received comparatively less attention. For example, Bosilovich and Schubert (2002) performed integrations from June through August. Additionally, Rasmusson (1967) and Mestas-Nuñez et al. (2007) analyzed and contrasted the winter and summer seasons. However, analyzing the spring season is important for several reasons: (1) it is a season typified by frequent Midwest flood events; (2) it is the primary season for tornadoes in the southern US affected by the availability of low-level moisture from the Gulf of Mexico; (3) rainfall anomalies in the spring create soil moisture conditions that can carry over to summer and either exacerbate or ameliorate the tendency for summer floods or drought conditions; and (4) it is observed that over the Mississippi River basin the moisture flux convergence is largest during spring and is associated with the northward flux from the Gulf of Mexico (Berbery and Rasmusson 1999).
The influence of the El Niño-southern oscillation (ENSO) phenomenon on the tropical North Atlantic has been well-documented. However, other teleconnections that are active during winter and spring months also play a role in the IAS. For example, one of the characteristics of the Pacific North American (PNA) oscillation is its anomaly center of sea level pressure straddling the southeastern part of the US (Feldstein 2002; Barnston and Livezey 1987; Wallace and Gutzler 1981). The PNA has also been documented to have a major influence in the region of the Ohio River Valley (Coleman and Rogers 2003) and the Mississippi River region (Roger and Coleman 2003). Also, the Pacific decadal oscillation (PDO) influences the precipitation variability in North America (McCabe et al. 2004; Mantua and Hare 2002). Nevertheless, the PNA and the PDO have been discussed less with respect to their impacts and relation to IAS climate variability.
Tornadic activity is also influenced by moisture fluxes from the Gulf of Mexico. A well-known region of tornadic activity is the “Tornado Alley”, comprised mainly of the states of Texas, Oklahoma, Kansas, Nebraska, eastern Colorado, and South Dakota with most of the activity occurring during May and June. However, other regions of the US are also prone to tornadic activity throughout the year (Brooks et al. 2003a), sometimes more prominent in late winter and early spring and with a different seasonal behavior when compared with Tornado Alley. For example, one of the major tornado outbreaks in the US occurred on April 3–4, 1974 in the region of the Mississippi and Ohio River basins outside of Tornado Alley (Verbout et al. 2006). Additionally, Trapp et al. (2005) found that tornado activity was strongly associated with squall lines and bow echoes (as opposed to cells) for the US states situated along a curved axis from Louisiana to Pennsylvania and that activity was higher in the first few months of the year (i.e., from January to April). Tornadic activity in a similar region was found by Cook and Schaefer (2008) to be susceptible to the influence of ENSO conditions in boreal winter.
The objective of the present study is to analyze the climate factors associated with the principal mode of interannual variability for the Intra-Americas low-level jet (IA-LLJ) in boreal spring and its relation to the precipitation and tornadoes of the Mississippi and Ohio River basins. We analyze the relation of the IA-LLJ variability with Pacific and Atlantic teleconnections with the goal of identifying the dominant regional mechanisms likely forced from outside the IAS.
2 The data and the approach
The 925-hPa Caribbean zonal wind and the Gulf of Mexico meridional wind are most highly and significantly anti-correlated in late winter and early spring (February, March, and April) and in September in both the ERA-40 and NARR data sets. This indicates that when the anomalous winds in the Gulf of Mexico are southerly, the anomalous winds in the Caribbean are most likely easterly. It is in these months that the moisture fluxes of the Caribbean would combine with the moisture fluxes of the Gulf of Mexico with greater impact on the US continent. This implies a greater moisturization of the low-level air mass due to the longer fetch over warm waters. The months with the lowest correlation in the NARR data set were July and August, although the ERA-40 data set showed a modest correlation for these summer months.
To analyze the main mode of variability of the IA-LLJ the corresponding time series (or PC1) was regressed on other atmospheric and oceanic fields and correlated with the teleconnection indexes. The ERA-40 data set was used for the following atmospheric parameters: sea level pressure (SLP), vertical velocity, geopotential height, divergence, air-sea fluxes, moisture fluxes, specific humidity and winds. For precipitation, the University of Delaware’s 0.5 × 0.5 degree data set was used (Willmott and Matsuura 2007); for sea surface temperature (SST), NOAA’s Extended Reconstructed SST data set (ERSST.v3) was used (Smith et al. 2008).
To analyze the relationship between the IA-LLJ and US precipitation and tornadoes we calculated two additional indexes. One index was for the area-average precipitation anomalies over the region of the lower Mississippi, Tennessee, and Ohio River basins. We refer to this index as the MORB precipitation index. The University of Delaware (Willmott and Matsuura 2007) precipitation data set was used to calculate the MORB precipitation index. The other index was for the monthly average of tornado counts for the region encompassing the MORB precipitation index. The data set used to calculate the tornado activity index was the Storm Prediction Center Tornado Database (McCarthy 2003; Schaefer and Edwards 1999). As Doswell et al. (2009) indicate, analyzing tornado data can be challenging due to issues associated with the temporal inconsistency of the tornado intensity Fujita scale (F-scale). For example, it is likely that pre-1975 tornadoes were rated higher (Verbout et al. 2006). Therefore, we proceeded with caution in creating the tornado activity index and accounted only for those tornadoes of significant strength.
To analyze the environmental conditions related to the tornado activity index, we used the covariance of wind and specific humidity data from the NCEP/NCAR Reanalysis (Kalnay et al. 1996). Mestas-Nuñez et al. (2005) showed the usefulness of the NCEP/NCAR Reanalysis moisture fluxes for the water budget analysis of the IAS. We also used wind shear and convective available potential energy (CAPE) data from NARR (Mesinger et al. 2006). CAPE is a measure of the amount of energy available for convection, and Brooks et al. (2003b) identified the CAPE and the wind shear as quantitative measures that serve to characterize environments conducive to severe storms.
The following teleconnection indexes were also used: the PNA index and the north Atlantic oscillation (NAO) index from NOAA’s Climate Prediction Center calculated according to the methodology of Barnston and Livezey (1987); the PDO index from Miller et al. 1994; and the tropical Pacific SST anomaly Niño indexes.
3 Principal mode of variability of IAS low-level winds
3.1 The atmospheric circulation variability
The main mode of variability for the March–April 925-hPa winds of the Intra-Americas region has a spatial pattern (EOF1) in which the Caribbean easterlies and the Gulf of Mexico southeasterlies change along the axis of their climatological flow (Fig. 4b). This first mode has more than 70% of the explained variance, and out of the first several loading vectors (EOFs) is the only one with a coherent intensification (or weakening) of the climatological flow (the IA-LLJ). The atmospheric flow from the Caribbean into the Gulf of Mexico through the Yucatan Channel weakened, for example, during the years of 1969, 1987, and 1993, while it strengthened for the years from 1973 through 1977 (Fig. 4b, c). In fact, a decadal shift in the late 1970s can be identified as the direction of the IA-LLJ PC1 anomalies is mostly positive before the early 1980s and mostly negative after the late 1970s.
Correlation of IA-LLJ PC1 and northern hemisphere teleconnection indexes for 1958–2001
North Atlantic oscillation
East/north Pacific pattern
Niño1 + 2
Pacific decadal oscillation
Pacific North American pattern
3.2 Sea surface temperature
In the region of the IAS, there is a cooling of the Caribbean Sea and a warming of the Gulf of Mexico associated with a strengthening of the IA-LLJ (Fig. 6b). The warming in the Gulf of Mexico and along the east coast of the US is associated with a decrease in heat loss from the ocean by turbulent and radiative heat fluxes (not presented), of which the latent heat flux is dominant followed by the sensible heat flux. This relation with latent and sensible heat fluxes is consistent with changes in the intensity of frontal passages associated with the anomalous pressure trough over the eastern seaboard (Eichler and Higgins 2006). The oceanic region off the southeastern US also shows positive SST anomalies as a result of the negative windstress curl anomalies. However, the cooling in the Caribbean Sea is associated predominantly with an increase in evaporation (negative latent heat flux anomalies) due to the strengthened northeast trades (Fig. 5c). Muñoz et al. (2009) have done an in-depth analyses of the IAS SST dipole during springtime.
A strengthening of the IA-LLJ is associated with a cooling of the tropical Pacific SST anomalies that serve as background for the influence of the PNA. The negative phase of the PNA is associated with cooling along the west coast of North America and warming farther west as observed in Fig. 6b. The cooling of the tropical Pacific is marginally significant, with a correlation between the Niño3 index and the IA-LLJ PC1 of 0.38. Similarly, the negative (cool) phase of the PDO is associated with a strengthening of the IA-LLJ (Table 1). This tropical Pacific cooling acts as a modulator by strengthening the tropical North Atlantic (TNA) trade winds, inducing latent heat flux anomalies that cool the TNA surface waters. Enfield and Mayer (1997) presented evidence of a lagged and positive correlation between the tropical Pacific conditions in winter and the tropical North Atlantic conditions a few months after in boreal spring. In this case, stronger easterlies induced an increase in evaporative heat loss and consequent cooling of the Caribbean SSTs.
4 Precipitation variability in the Mississippi and Ohio River basins
We compared the IAS moisture flux anomalies related to the MORB precipitation index with those related to the precipitation anomalies of the Great Plains during March–April. To create the Great Plains index of March–April precipitation anomalies, we used the area delimited by Ruiz-Barradas and Nigam (2005) as the area of the Great Plains (i.e., 35–45°N, 90–100°W) (purple square in Fig. 9a). In contrast with the MORB index, the Great Plains precipitation index for March–April shows precipitation anomalies centered over 93°W and 38°N that are weaker than those of the MORB index. The variability of the Great Plains index in March–April is centered in the southeast quadrant of the Great Plains index area. Positive Great Plains precipitation anomalies in March–April are related to stronger moisture fluxes from the Gulf of Mexico but not (statistically significant) from the Caribbean Sea. The moisture flux anomalies extend from the northern area of the central US southward towards Mexico. The anticyclonic circulation centered over the southeastern US is less evident, while a stronger cyclonic flow centered over the states of California and Arizona emerges. The zones of moisture flux convergence are also different, as the stronger Great Plains index is associated with moisture flux convergence over the north-central US and moisture flux divergence to the west of California and the Baja California peninsula. For both precipitation indexes, the moisture fluxes from the Gulf of Mexico are important. However, the associated circulation anomalies are distinct for each case. The Caribbean moisture fluxes are less related to the circulation anomalies that affect the Great Plains region (used for the index) in boreal spring.
The SLP and SST anomalies associated with the MORB precipitation index are highly similar to those related to the IA-LLJ variability (Fig. 6). The SLP and SST patterns (not presented) highlight significant anomalies in the midlatitudes of the North Pacific. A strong positive zonal gradient of SLP from the region of Texas to the east of the US is an indication of the strengthening of the meridional component of the wind at the lower levels. The region of the Gulf of Mexico and to the east of the US also shows warm anomalies associated with an increase in precipitation in the MORB region. These warm anomalies are expected as the precipitation bands associated with storm tracks move north during negative PNA and La Niña events; this results in decreased cloudiness over the Gulf region and decreased heat loss from turbulent fluxes. In fact, the teleconnection index most highly (anti-) correlated with the MORB index in March is the PNA (−0.43). The PNA, however, is not significantly correlated with the MORB index in April. The Niño3.4 index is the teleconnection index most highly correlated with the MORB index in April. A correlation of 0.40 between the February Niño3.4 and April MORB indexes is perhaps an indication of the superposition of the lead influence of the ENSO and the concurrent influence of the PNA.
5 Tornado index
We further explored the relation between the IA-LLJ and extreme events in the region encompassing part of the lower Mississippi, Tennessee, and Ohio River basins in March–April by analyzing the interannual variability of tornadic activity. Tornado data from the Storm Prediction Center (McCarthy 2003; Schaefer and Edwards 1999) were used to create a monthly index of tornado counts for 1950–2006 for the following nine states: Alabama, Arkansas, Illinois, Indiana, Kentucky, Louisiana, Mississippi, Missouri and Tennessee. The tornadoes counted were category F2 and stronger (i.e., significant tornadoes) on the F-scale for tornado intensity. The tornadoes were grouped into a regional monthly tornado count index (TCI) for the nine states. As our interest was for how climate variability relates to tornadic activity, we smoothed the monthly time series with a 1-2-1-month running mean and rounded the values to the nearest integer. For example, the TCI value of March 1973 is substituted by [TCI(February 1973) + 2*TCI(March 1973) + TCI(April 1973)]/4. From the smoothed time series, the tornado count for April 1974 was the largest. The monthly median (1951–2006) for the smoothed time series was rounded to the nearest integer and removed from each respective month. The median was chosen as the measure of central tendency since the tornado index appears to have a skewed distribution. Nevertheless, the values for the year 1974 were not accounted for in the calculation of the median. They were excluded from the time series because 1974 was the most extreme year and would have affected the statistics.
6 Climatic conditions related to the tornado index
We regressed the tornado index to the low-level (925-hPa) and mid-level (500-hPa) winds and specific humidity. The regression coefficients (not presented) make evident that the pattern of low-level winds associated with tornadic storms is very similar to the pattern associated with the main mode of variability for the IA-LLJ (Fig. 5). An intensification of the Gulf of Mexico southerly winds and the Caribbean easterly winds occurred and, in the North Pacific, there were significant anomalies in the midlatitudes. It is also evident that the low-level moisture flux from the Gulf of Mexico is important in providing the moisture relevant to the development of tornadic storms in the region of study. At the mid-levels (e.g., 500-hPa), the pattern of the winds was one with strong anomalies coming from the southwest over the Rocky Mountains. Although there are positive anomalies of specific humidity associated with these winds, they are of very low strength.
Rank correlation coefficients between the regional tornado index and teleconnection indexes for 1950–2007
PDO January (−0.43)
N34 December (−0.31)
PDO February (−0.41)
N34 January (−0.39)
PNA March (−0.46)
N34 March (−0.41)
PNA March (−0.33)
N34 April (−0.20)
NAO May (−0.21)
N34 May (−0.13)
7 Summary and conclusions
The IA-LLJ forms a corridor that transports moisture into North America from the Gulf of Mexico and the Caribbean Sea. However, the focus of previous studies has been limited to understanding the winter and summer IAS moisture transport. The objectives of this study were to diagnose the dynamics of the IA-LLJ in boreal spring, its remote forcings, and its impacts.
We found that in recent decades (1980s to mid-2000s) the Caribbean and Gulf of Mexico low-level winds have been highly correlated during March–April. The variability of the IAS 925-hPa wind anomalies in March and April was analyzed here by principal component analysis (PCA) using data for 1958–2001. The PCA reveals a main mode of variability in which the climatological 925-hPa wind weakens or intensifies in unison, indicating a fluctuation of the IA-LLJ and its moisture transport into the US east of the Rocky Mountains. The main teleconnection associated with the IA-LLJ variability is the PNA. Prediction of the IA-LLJ variability may, therefore, be promising given that Hamill et al. (2006) found the PNA to be one of the three most predictable patterns with a 10-day lead.
The strengthening of the IA-LLJ increases the source of moisture from the IAS, having a positive impact on the precipitation in the Mississippi, Tennessee, and Ohio Rivers basins. This moisture influx at low levels not only impacts the precipitation but also the tornadic activity in the region centered at the junction of the Ohio and Mississippi rivers. A regional index of tornado count (TCI) variability was constructed and found to have similar influences as those factors affecting the IA-LLJ variability. Other associated factors that influenced the TCI variability were: (1) wet transients of zonal and meridional moisture fluxes at the low levels; (2) dry transients of zonal moisture fluxes at the levels between 600 and 700 hPa; (3) southwesterly wind shear between the 500-hPa winds and 10-m winds; (4) positive CAPE anomalies over the southeastern gulf states; and (5) positive precipitation anomalies. Yet, the relation of tornadic activity in this region with preceding precipitation is unclear; although Shepherd et al. (2009) found that springtime tornadic activity for the state of Georgia had a tendency to be lower when drought conditions persisted in the preceding fall and winter.
The variability of the three indexes developed for this study (i.e., IA-LLJ index, precipitation index, and tornado index) has a strong relationship with the variability of the PNA and the PDO. Based on our analyses, we conclude that the negative phase of the PNA during boreal spring is related to an intensification of the IA-LLJ, which provides greater moisture to the region of the Mississippi and Ohio River basins, leading to increased precipitation and greater tornadic activity. The significant relation with the PDO is also evident in the decadal shift of the three indexes in the late 1970s, consistent with the behavior observed herein. That is to say, the weakening of the springtime IA-LLJ during the 1980s and 1990s led to weaker moisture influx from the IAS, thereby manifesting as a decadal shift (a decrease) in precipitation and tornadic activity in the region of the Mississippi and Ohio River basins for the same period.
In this study, we have presented evidence that the IA-LLJ (and its associated moisture fluxes) is one dynamic feature by which the PNA, the PDO, and ENSO influence precipitation and tornado variability in the region of the lower Mississippi, Tennessee, and Ohio River basins. This dynamic feature is important with respect to seasonal hydrological and tornadic activity forecasts.
This research was carried out in part under the auspices of the Cooperative Institute for Marine and Atmospheric Studies (CIMAS), a cooperative institute of the University of Miami and the National Oceanic and Atmospheric Administration, cooperative agreement #NA17RJ1226. This work was also supported by base funding of NOAA’s Atlantic Oceanographic and Meteorological Laboratory (AOML). The findings and conclusions in this report are those of the author(s) and do not necessarily represent the views of the funding agency. We thank the reviewers’ comments and suggestions. Gail Derr gave some editorial comments. The ERA-40 data used in this study were provided by ECMWF and NCAR.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- Amador JA (2008) The Intra-Americas Sea low-level jet: overview and future research: In: Trends and directions in climate research. Ann NY Acad Sci 1146:153–188Google Scholar
- Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M, Saha S, White G, Woollen J, Zhu Y, Leetmaa A, Reynolds R, Chelliah M, Ebisuzaki W, Higgins W, Janowiak J, Mo K, Ropelewski C, Wang J, Jenne R, Joseph D (1996) The NCEP/NCAR 40-year reanalysis project. Bull Am Meteorol Soc 77:437–471CrossRefGoogle Scholar
- McCarthy DW (2003) NWS tornado surveys and the impact on the national tornado database. Preprints, First Symposium on F-scale and Severe Weather Damage Assessment, Long Beach, CA. Am Meteorol Soc., CD-ROM, 3.2Google Scholar
- Miller AJ, Cayan DR, Barnett TP, Graham NE, Oberhuber JM (1994) The 1976–1977 climate shift of the Pacific Ocean. Oceanogr 7:21–26Google Scholar
- Muñoz E, Wang C, Enfield D (2009) The Intra-Americas Sea springtime surface temperature anomaly dipole as fingerprint of remote influences. J Clim (in press)Google Scholar
- Roger JC, Coleman JSM (2003) Interactions between the Atlantic multidecadal oscillation, El Niño/La Niña, and the PNA in winter Mississippi Valley stream flow. Geophys Res Lett. doi:10.1029/2003GL017216
- Schaefer JT, Edwards R (1999) The SPC Tornado/Severe Thunderstorm Database. Preprints, 11th conference applied climatology, DallasGoogle Scholar
- Wilks DS (2006) Principal component (EOF) analysis. In: Statistical methods in the atmospheric sciences, 2nd edn. Elsevier, London, pp 463–508Google Scholar
- Willmott CJ, Matsuura K (2007) Terrestrial precipitation: 1900–2006 Gridded Monthly Time Series, http://climate.geog.udel.edu/~climate/html_pages/Global_ts_2007/README.global.p_ts_2007.html. Accessed 2008