A Climatology of Heavy Snowfalls in Northwest Missouri: Long Term Trends and Interannual Variability.

Cynthia I. Berger [1,3]

Anthony R. Lupo [1][*]

Peter Browning [2]

Micheal Bodner [2]

Christopher C. Rayburn [1]

Matthew D. Chambers [1]

Abstract

One of the more difficult forecasting challenges in northwest Missouri (NWMO) is the arrival of heavy snowfalls, which can occur frequently during the cold season and usually in association with synoptic-scale waves. The goal of this study is to develop a 50-year snowfall climatology using a dense station observation network which covers northwest and part of central Missouri. This includes a study of long-term trends and interannual variability in order to determine the extent to which sea surface temperature variations in the Pacific Ocean basin correlate with interannual variability in snowfall occurrence for northwest Missouri. The region of study includes the county warning area for which the National Weather Service Forecast Office (WSFO) in Pleasant Hill, Missouri, is responsible. This study was performed using station data that is available through the Missouri Climate Center. The specific objectives were: 1) to compile a snowfall database which can be used for reference by the WSFO in Pleasant Hill, 2) to examine the statistics, such as frequency of snowfall occurrence, and 3) to examine long term trends and interannual variability in snowfall occurrence which may relate to the El Nino and Southern Oscillation (ENSO) and the North Pacific Oscillation (NPO).

Initial results demonstrate that northwest Missouri can expect about three heavy (6 to 10 inches total) and/or extreme (total of 10 inches or greater) snowfall events during the winter season. While no significant trend in heavy snowfall occurrence has been noted, there has been a reduction in the number of extreme events over the region. Also, fewer snowfall events were noted during El Nino years, while more heavy snowfall events occurred during years classified as "neutral." However, a closer examination of the results demonstrated that El Nino/La Nina-related variability in snowfall occurrence is superimposed on longer NPO variability. These results were also broken down by season and demonstrated ENSO and NPO related variability.

1. Introduction

In recent years, the issue of climate change and the dynamic mechanisms leading to climate change on global and regional scales has garnered the interest of the research community, as well as the interest of the general public (e.g., IPCC 1996). Various mechanisms have been examined in order to explain long and short-term global climate variations, including the effect of increasing CO2 concentrations, which have ostensibly been linked to human activities (e.g., IPCC 1996; Singer 1997). Climate fluctuations due to natural variability inherent in the climate system have also been examined extensively. In particular, shorter-term oscillations, or interannual variability in local climates have been linked to coupled ocean-atmosphere phenomena such as the El Nino and Southern Oscillation (ENSO), the North Atlantic Oscillation (NAO) (e.g., Hurrell 1995; Gray 1998), and most recently, the North Pacific Oscillation (NPO) (or the Pacific Decadal Oscillation - PDO; e.g., Gershunov and Barnett 1998). While the controll ing mechanisms for the NPO are, as of yet, poorly understood (e.g., Gershunov and Barnett 1998), the NAO is believed to be linked to decadal and century-scale variability in the North Atlantic thermohaline circulation (e.g., Wunch 1992; Schmitz 1996; Gray 1998). The controlling mechanisms driving El Nino are also not completely understood. It is widely believed, however, that complex ocean-atmosphere interactions may generate easterly propagating oceanic Kelvin waves which are responsible for depressing the east Pacific thermocline and transporting warmer waters from the western Pacific to the eastern Pacific (McPhaden 1999).

The atmospheric signal from these SST phenomena results from the forced response of the atmosphere to sea surface temperature (SST) variations. SST anomalies have been shown to influence the mean structure of mid-latitude atmospheric circulations on time scales from seasons (McPhaden 1999) to a few years (e.g., ENSO-related variability) to decades (e.g., NPO-related variability; Gershunov and Barnett 1998). Many studies have demonstrated, using observational data and modeling results, the importance of SST variations in forcing an atmospheric response on short and long temporal and spatial scales (e.g., Namais 1982; Hoskins et al. 1983; Kung et al. 1990, 1992, 1993; Nakamura et al. 1997; Lau 1997; Lupo and Bosart 1999).

The influence of SST anomalies are not limited to the ocean basin of their occurrence, as is well-known in the case of the El Nino phenomenon (e.g., Gray 1984; O'Brien et al. 1996; Bove et al. 1998), and demonstrated recently for the NPO (Minobe 1997; Gershanov and Barnett 1998). They influence the local climatological characteristics of atmospheric variables downstream as well (e.g., NAO influence on European climate: Hurrell 1995; Wunsch 1999, or ENSO on United States climate) via their impact on well-known Northern Hemispheric teleconnection patterns (e.g., Wallace and Gutzler 1981; Namais 1982; Blackmon et al. 1984; Mo and Livezey 1986; Hurrell 1996). Recently, Karl and Knight (1998) have demonstrated ENSO-related variability in United States temperatures and precipitation, while Livezey et al. (1997), Hu et al. (1998), and Berger et al. (1999) have noted ENSO-related variability in Midwestern precipitation. Hu et al. (1998) also showed variability in Mid-western precipitation on longer time scales as wel l as a general upward trend in precipitation amounts. Karl and Knight (1998) also noted these general trends for heavy precipitation events. Karl et al. (1993) and Karl et al. (1996) have shown long-term trends in United States temperatures as well.

One of the more difficult forecasting challenges for northwest Missouri (NWMO) this area is the arrival of heavy snowfalls, which can occur frequently during the cold season. Heavy snow can affect communities in many ways, including the slowing down or halting of air and ground traffic flow, which in turn can have a large economic impact on the region. The timing of heavy snowfalls can also be a major factor on their economic impact. An early heavy snowfall event in October could adversely affect the late season harvest. A late heavy snowfall event in the spring could also affect the farming industry by delaying the planting of crops and saturating the ground. In addition, these snowfall events can also contribute to seasonal flooding problems.

Heavy snowfalls are events that typically occur in association with synoptic-scale waves. However, these snowfalls often occur on temporal and spatial scales more consistent with those of mesoscale phenomena. Therefore, identifying the common climatic, synoptic and mesoscale patterns that produce heavy snowfall is essential for the improvement of heavy snowfall forecasting. There have been a few recent studies that discuss the climatological characteristics of snowfalls in NWMO (e.g., Metze et al. 1998; Berger et al. 1999). However, these studies do not provide the necessary detail needed to comprehensively understand the climatological behavior and interannual variability of snowfall occurrence in the NWMO region.

In this study, a comprehensive climatology of heavy snowfall events is presented using a dense station observation network covering NWMO and part of central Missouri. The region includes the county warning area of the National Weather Service Forecast Office in Pleasant Hill, Missouri, which was a partner in this study. Statistics were examined in order to find long and short-term trends and study the seasonal and intraseasonal variability, as well as interannual variability in snowfall occurrences and their relation to ENSO and NPO variability. Thus, the objective of this study is to gain a thorough understanding of the climatological character, including the interannual variability, of heavy snowfalls in NWMO.

2. Data and Methodology

a. Data Sources

Snowfall data were taken from the Missouri Climatological Data (MCD), which are archived in periodical format at the Missouri Climate Center (MCC) on the University of Missouri-Columbia campus, and made available through the MCC. These archives include a variety of climatological parameters including temperature and precipitation records at weather observation stations throughout Missouri. Hourly Precipitation Data (a companion publication with the MCD), the daily 500 hPa, and surface weather maps archived in atlas format and on microfilm were also examined in order to verify whether multiple-day snowfalls were single snowfall events. These data can also be found at the MCC. The daily weather map series is a weekly publication available through the National Center for Environmental Prediction (NCEP).

b. Definitions

A 50-year period was chosen for this study starting with the 1949-1950 snowfall season. This period was chosen because it was sufficiently long enough to describe significant interannual variability, including a complete cycle of the North Pacific Oscillation. A longer time period was not chosen because records are less reliable prior to the 1940's. A snowfall season is defined as starting on 1 October and running through 30 April. The data were stratified in this manner as opposed to the calendar year in order to follow more closely the annual cycle in occurrence of snowfall events and the definition of El Nino/La Nina year used in this study. This type of seasonal categorization has been used in the research of other atmospheric phenomena such as blocking anticyclones (e.g., Quiroz 1987; Lupo and Smith 1995). The annual snowfall season was then subdivided into calendar seasons, with fall season snowfalls occurring in the months of October and November, winter season snowfalls from December through February, and spring season snowfalls occurring in March and April. Snowfall events were categorized as moderate (3 to [less than] 6 inches), heavy (6 to [less than] 10 inches) and extreme (10+ inches). Events were not limited to, for example, a 24 h time period from 0000 UTC to 0000 UTC (e.g., Metze et al. 1998; Glass 1998), but instead were classified as any continuous snowfall occurring in association with a larger scale feature. For instance, a slow moving weather system could produce snow in NWMO (see Fig. 1 for a map of the region) over a time period covering 2 to 3 calendar days, and for the purpose of this study it was considered one event. In the summer, isolated rainfalls would need to be eliminated in a study such as this because of their non-organized nature (e.g., Glass 1998), however, winter storms in NWMO tend to be more widespread spatially. In categorizing events, it was required that only one station in our region of study report a snowfall amount in a particular category, even if all others recorded less snowfall. Thus, if one station recorded 10.5 inches of snow over the course of one event while other stations recorded less than 10 inches, the event was categorized as extreme because of the storm potential to produce extreme snowfall amounts.

ENSO was defined in this study according to the Japan Meteorological Agency (JMA) ENSO Index. A complete description can be found on the Center for Ocean and Atmospheric Prediction Studies (COAPS) [1]. The index classifies years according to sea surface temperature (SST) anomaly thresholds. Three phases are described: El Nino (EN) is used for warm anomalies, La Nina (LN) is used for cool anomalies, or neutral (NEU) is used when no significant anomaly is found. SST anomalies derived from 5 month running means are used covering the tropical Pacific from 4[degrees]S-4[degrees]N, 150[degrees]W-90[degrees]W. A year is classified as El Nino (La Nina) if the index values are 0.5[degrees]C (-0.5[degrees]C) or higher for 6 consecutive months, of which the six months must begin before October of the previous year and continue through December. Values between -0.5[degrees]C and 0.5[degrees]C (but not inclusive) would be considered neutral. The 'El Nino' year is defined as starting in October of the previous year and continuing through the following September. A list of the years in this study separated by ENSO phase is found in Table 1, which is also borrowed from [COAPS.sup.1].

The North Pacific Oscillation, or NPO, is a long-term SST oscillation occurring over a 50 to 70 year time period (Minobe 1997) within the Pacific Ocean basin. This oscillation is also called the Pacific Decadal Oscillation (PDO). As defined by Gershunov and Barnett (1998), the high (positive) phase of NPO is characterized by an anomalously deep Aleutian low. Cold western and central North Pacific waters, warm eastern Pacific coastal waters and tropical Pacific waters also characterize this phase of the NPO. We referred to this phase as NPO1. The reverse conditions characterize the low (negative) phase of NPO and we referred to these conditions as NPO2. Table 2 defines the period of record for each phase of the NPO since period 1933. These are based on the findings of Gershanov and Barnett (1998) and described by Kerr (1999).

c. Procedures

After classifying years as El Nino, La Nina, and neutral, simple means were calculated to examine and compare decadal-, ENSO-, and NPO- related variability. Comparisons of means were examined using a two-sided "simple standardized test statistic" ([z.sup.*]; Neter et al. 1988, pp. 310 - 366). Means for the entire 50-year period sample served as the "expected" frequency of occurrence. Since the distributions were not known beforehand, the two-tailed test was used since it results in a more stringent criterion for significance. In order to find long-term trends over the length of the data set, simple regression lines were constructed. The significance of trends were examined using an analysis of variance approach (ANOVA) (e.g., F-test; see Neter et al. 1988, Ch. 19). All statistical tests assumed the null hypothesis, or that there is no a priori relationship between the two variables being tested.

Additionally, in order to find and confirm the presence of longer-term variability in our data, Fast Fourier transforms (FFTs) were applied to our time series to construct periodograms, and these time series were examined by removing the mean. The FFTs were applied to discrete time series, and in order to extract meaningful information, the 99% confidence level was plotted against a white noise background continuum in order to determine which "peaks" were statistically significant. A more complete description of FFTs and programming techniques can be found in Press et al. (1998).

3. Climatological Analysis

a. Seasonal Variations and Decadal Trends

The MCD data were examined over the period beginning with the 1949-1950 snowfall season and ending with the 1997-1998 season. Snowfall events occurred between the months of October and April and totaled 391 for the period (Table 3), resulting in an average of 8 events per year over the region of study. As expected, most of the events were categorized as moderate (238 events or 61%), with few extreme events (29 or 7%). While the majority of these events occurred in the winter season (December-February: 275 events), there were more events that occurred in the spring months (83 events or 21%) than in the fall (33 events or 8%; Table 3). A monthly plot of snowfalls (Fig. 2a) shows that the distribution was nearly normal with a peak occurrence of 2 events occurring annually in January and February. Figures 2b, 2c and 2d break down the monthly distribution into snowfall categories. Moderate snowfall events have a distinct peak from December to February. Heavy snowfalls tended to be skewed slightly toward a peak in February, and extreme snowfalls had a peak in January.

Interdecadal variability was also examined according to season (Tables 4a, 4b, and 4c). A slight upward trend was found in the total number of fall snowfalls, with a higher frequency of fall snowfalls in the last 3 decades, especially when comparing moderate plus heavy events. Also, for fall season events (Table 4a), there was no trend for extreme or heavy snowfalls. Then, the fall season increase was due to an increase in moderate events alone. Winter snowfalls showed a slight downward trend in the total occurrence of events (Table 4b), though the trend was strongest over the late 1980's and 1990's. There was no trend in moderate or heavy events, but a downward trend was evident in the extreme category. Heavy (moderate) snowfalls showed a peak occurrence in the 1970's (1950's). The examination of heavy plus extreme winter events revealed that the majority of events were found in the 60's and 70's, with a downward trend thereafter. A downward trend was found in the total spring snowfalls, as well as within e ach category (Table 4c). These trends were especially evident when comparing moderate plus heavy, and heavy plus extreme events, and these trends mimicked the total downward trend. Only the downward trend in total and heavy spring snowfall events proved to be significant at the 95% level when using the F-test.

b. Long Term Trends

The total number of annual events with respect to time is plotted in Fig. 3, and the regression analysis shows that there is a slight downward trend in the total number of snowfall events occurring annually. This trend did not prove to be significant at the 95% level using the F-test. Figures 5a, 5b, and 5c show the distribution of snowfalls divided into categories. The trend is also downward, but to varying degrees in each separate category, though none of these trends proved to be significant at the 95% level. These results are occurring despite the general increase in precipitation in the Midwestern United States over the last 30 years (Hu et al. 1998; Karl and Knight 1998) and decreasing temperatures in the southern Plains and southeastern United States (Karl et al. 1993; Fig. 4). However, fewer snowfall events would agree with the general upward trend in global and Northern Hemisphere average temperatures (TPCC 1996; Bell and Halpert 1998; Bell et al. 1999). A partial explanation of this contradiction, however, could be revealed by examining the synoptic trends (cyclone frequencies and storm track variations) and dynamic structure of individual snowfall events themselves. For example, the studies of Zishka and Smith (1980) show downward trends in the number of winter and summer season cyclones over North America from the 1950's through the late 1970's. Also, Key and Chan (1999) find downward trends in mid-latitude cyclones through the late 1990's. Further studies of these issues are being continued, especially with regard to the dynamic structure of individual cyclones.

Examining the decadal variability of the entire sample shows that the first three decades (1950's, 1960's and 1970's) averaged 8.3 snowfall events per year, while the last two decades (1980's and 1990's) have averaged 7.3 events per year (Table 5). The trend is most severe for the extreme category. The first three decades averaged 7 extreme events per decade, while in the past two decades the average has been 4.7 events per decade. This downward trend is counter to the increase in frequency of heavier rainfall events (Karl and Knight 1998). There was a downward trend in the number of heavy events (2.8 events per year in the first three decades vs. 2.1 events per year in the final two), and no significant trend in the number of moderate events. Also, when moderate and heavy events are examined together, the trend is downward, but not significant. Even though there is a slight downward trend in the overall frequency of snowfall events examined here, the interdecadal variability mirrors that found by Karl and K night (1998) for the frequency of the upper 10-percentile of daily precipitation events. A more detailed examination of this variability is discussed in section 4.

4. ENSO- and NPO-related Variability

a. E1 Nino vs. La Nina Years

There was evidence of interannual variability not only in the total number of events, but within each category as well (Fig. 6). A simple FFT analysis demonstrated that there is significant variability on the 25-year cycle. This variability is longer than the 20 and 12-year variability suggested by Hu et al. (1998), however this result might be a function of the length of our data set and the methodology chosen. Since we chose a 50-year period and removed the mean (Fig. 6), this peak may be a sub-harmonic of NPO related variability. Note that the NPO completed on cycle over virtually the same period of time chosen for our analysis (Table 2). Our analysis also shows other peaks (at 4, 2.5, and 2-year cycles), which demonstrate variability consistent with the time scale of ENSO variability (2 to 7 years).

Table 6 displays the number of events vs. E1 Nino/La Nina phase. A majority of snowfall events occurred during the NEU phase, as was expected since our sample includes 26 NEU, 11 LN, and 13 EN years. There were a similar number of total overall events when comparing EN and LN years. However, during EN years, 7.3 events occurred on the average, compared to 8.3 events for LN and 8.2 for NEU years. While there was no difference in the mean number of extreme events per year for any category, there were more moderate events in LN years compared to NEU and EN years, and more moderate plus heavy events in LN and NEU years compared to EN years. Thus, the more frequent occurrence of E1 Nino events in the past two decades (Table 1, 1980's and 1990's) may partially explain the general decrease in overall snowfall events shown in section 3. None of these results were significant at the 95% or 99% level, If any results in this section prove significant at these levels, they will be noted in the text.

A more detailed analysis of NEU years only was performed (not shown). In neutral years following or preceding E1 Nino (La Nina), the means were similar to those of La Nina (E1 Nino) years. However, in NEU years that served as a transition from one phase to the other, the annual snowfall occurrences were more similar to those during EN years. In NEU years where there was no phase change (NEU years bookended by EN or LN years), the average occurrence of snowfall events was similar to that of LN years.

ENSO variations were also broken down by season. During the fall months (Table 7a) it was found that more snows occurred during EN years in the total sample. This is true across each category, with the exception of the extreme category, which was previously noted to have few occurrences during EN years. This agrees with the typical observation that falls in the Midwest tend to be cooler and wetter than normal for EN years [1]. Fewer snowfall events were found in EN winter months (Table 7b), which was evident across all snowfall categories. The spring results (Table 7c) were similar to the overall results, showing fewer snowfalls in EN years compared to LN and NEU years with the exception of the extreme category.

b. Variability With Respect to the NPO

More attention has been focused lately on longer-term oceanic oscillations that impact atmospheric circulations such as the North Atlantic Oscillation, and more recently the North Pacific Oscillation. Gershunov and Barnett (1998) found a correlation between NPO phase and the intensity of ENSO as it affects the atmospheric climatological flow regimes over the United States. They find that the NPO serves to either enhance or weaken the ENSO phenomenon, and thus the influence of the ENSO phenomenon, depending on the NPO phase. During NPO 1, the intensity of E1 Nino and its impacts on North American atmospheric climatological flow regimes tends to be greater, with a less intense La Nina related impact. The opposite is true for NPO2, which is indicative of stronger La Nina and weaker E1 Nino events. Thus, during NPO2, E1 Nino has less impact on typical North American circulation features.

Snowfall variability with respect to ENSO phase was examined in conjunction with the NPO. The NPO2 showed little ENSO-related variability in overall events when stratified by ENSO years (Table 8). During the NPO2 neutral years, however, fewer average occurrences were found in the moderate category, while a higher average occurrence was found within the heavy events, when compared to NPO2 EN or LN years. The occurrence of extreme events was less in NPO2 EN years when compared to LN and NEU years, but this result was not significant.

However, key differences were found when examining NPO1 (Table 9). LN and neutral years experienced more snowfall events both overall and within the moderate and heavy categories. La Nina (only one year in NPO1) and NEU years combined had similar average occurrences of snowfalls (8.4 per year), in NPO1 as was found for NPO2. However, fewer snowfall events were found during E1 Nina phase years. E1 Nino years showed an average of only 5.8 snowfalls per year, a result significant at the 80% confidence level (90% when considering NPO1 or NPO2 as the observed mean and testing versus each other). Thus, the reduced number of snowfalls in the warm phase of ENSO can be related to the enhancement of E1 Nino-related variability over North America by the NPO1. Categorically, moderate events exhibited the largest decline from LN plus NEU to warm phase. The decline in the total sample by percentage. The occurrence of extreme events is not consistent with the other categories, with more extreme events being found in EN yea rs during the NPO1 phase. When examining LN plus NEU together, the average occurrence is still less than that of EN years. The analysis shows that the character of ENSO-related variability over North America changes for NPO1, in that fewer LN years are found during this phase (Table 9). Also, it would seem that ENSO-related variability in the total sample is confined to, and reflects, ENSO-related variability in NPO1. This is confirmed by the fact that there was little ENSO-related variability in NPO2 as compared to NPO1.

The data in Tables 8 and 9 were also stratified by season in order to examine NPO and ENSO-related variability on this scale. Fall events displayed no significant ENSO-related variability in either NPO phase, especially in NPO1 (Table 10). Also, more snowfalls were found in NPO1 years, especially in the moderate category, a result consistent with those in section 3. that showed the increase in fall snowfalls during EN years. Winter season snowfalls varied little between NPO phase overall when examining individual snowfall categories. However, when ENSO phase was considered, the NPO variation found was similar to that of the total sample, and fewer events occurred during EN years (Table 11) in NPO1. During NPO2, the overall occurrence of snowfall events varied little, except there were more events that occurred during EN years. However, this result was not statistically significant. Additionally, moderate events occurred more (less) frequently in EN (NEU) years. More extreme events were found in LN and NEU ye ars.

The results for spring months are opposite that of fall months (Table 12). More snowfall events were found in NPO2 years than NPO1 years, especially in the moderate and heavy category. Little significant ENSO-related variability was found in NPO2 years, with the exception of extreme events, which were more common in NEU and EN years. During NPO1 years, only one snowfall was recorded during the LN year. The majority of snowfalls occurred in NPO1 NEU years, but the occurrence of spring snowfalls was similar for EN years and NEU years.

5. Summary and Conclusions

Heavy snowfalls are a phenomenon that can greatly affect a community. Understanding the long-term climatological characteristics of heavy snowfalls, including the climatological dynamics of the flow regimes, in northwest Missouri can aid in the forecasting process, which in turn will provide more accurate weather forecasts to the general public. In this study, several climatological trends and some important characteristics of long and short-term interannual variability for snowfall occurrence in northwestern Missouri have been identified using simple methodologies. A companion study is being carried out which will categorize these snowfall events by flow regime and examine their dynamic characteristics.

When examining seasonal variations, it was shown that the majority of snowfall events occurred in the winter months of December, January and February. More events were found in spring months than in fall months. There was also a trend towards more fall and fewer spring snowfall events. Only trends in the total and heavy spring snowfalls were found to be significant at the 95% level. However, in section 4 these trends were shown to be related to ENSO-scale interannual variability.

Overall, there was a downward trend in snowfall events. This result seems to contradict the findings of other studies showing an increase in Midwest precipitation and a decrease in temperatures in the region of study. This downward trend in snowfall occurrence does, however, agree with the general upward trend in global temperatures. A full explanation of this contradiction may be answered through a study of the synoptic trends (cyclone frequencies and/or storm track variations) and dynamic characteristics of individual snowfall events. An examination of snowfall occurrence by decade revealed that more snowfall events occurred during the first three decades of this study, and that the number of events has decreased in the last two decades. This trend was most severe in the extreme category. Again, none of these trends were found to be significant at the 95% level.

El Nino and Southern Oscillation has been the topic of many studies in recent years. ENSO variability of heavy NWMO snowfall events was examined in this study as well. It was found that there were more snowfalls in LN and NEU years than during EN years. This difference was noted primarily between the moderate and moderate plus heavy events. Seasonal variability was examined as well. During El Nino years, more snowfall events were found to occur in fall months. This corresponds to the common EN observation of cooler and wetter falls in the Midwest1. However, the decreasing trend in overall snowfall events over the 50-year period could partially be explained by the more frequent occurrence of El Nino years during the past two decades. An examination of the occurrence of neutral events shows that the number of snowfall events were similar to La Nina (El Nino) events following El Nino (La Nina) years.

Examining another long-term oscillation, the North Pacific Oscillation, has been of more recent interest. Other studies have shown that the positive phase of NPO, or NPO1, tends to enhance the effects of El Nino, and the negative phase, NPO2, tends to enhance La Ninas. In our study it was found that there was little variation in the total number of snowfall occurrences between NPO1 and NPO2. However, it was found that there was a significant difference between ENSO-related variability between each phase of the NPO. Thus, longer-term NPO-related variability was superimposed on smaller scale ENSO-related variability. More specifically, it was found there was little ENSO-related variability during NPO2. But during NPO1, there were significantly (at the 90% level) fewer snowfall events during El Nino years. Thus, this result demonstrates that the tendency for more (less) snowfalls to occur during La Nina (El Nino) years may not apply to an entire data set for a certain climatological parameter, but may also depe nd on interaction with other longer-term oscillations as well (e.g., Gershunov and Barnett, 1998). Also, while few of these results were found to be statistically significant, the results found here are consistent with changes in the basic atmospheric flow regime over North America as reflected by teleconnection patterns. The PNA pattern over North America (e.g., Wallace and Gutzler 1981) varies not only with El Nino phase, but, as shown by Gershanov and Barnett (1998), these teleconnections can be enhanced or weakened during NPO phase as well.

6. Acknowledgements

The authors would like to thank Dr. Stephen Mudrick, Dr. Qi Hu, and two anonymous reviewers for their insightful comments regarding this work. Special thanks also go to Rick Clawson for drafting some of the figures. This research was supported by the University Corporation for Atmospheric Research (UCAR) Cooperative program for Operational Meteorological Education and Training (COMET) Outreach Program under award # 98115921.

7. References

Bell, G.D., and M.S. Halpert, 1998: Climate assessment for 1997. Bull. Amer. Meteor Soc., 79, S1-50.

Bell, G.D., M.S. Halpert, C.F. Ropelowski, V.E. Kousky, A.V. Douglas, R.C. Schuell, and M.E. Gelman, 1999: Climate assessment for 1998. Bull. Amer. Meteor Soc., 80, S1-47.

Berger, C.L., A.R. Lupo, P. Browning, C.C. Rayburn, M.D. Chambers, and M. Bodner, 1999: The climatology of heavy snowfall events in northwest Missouri. Preprints of the 11th Conference on Applied Climatology. 10-15 January, 1999, Dallas TX.

Blackmon, M.L., Y.H. Lee, and J.M. Wallace, 1984: Horizontal structure of 500 mb height fluctuations with long, intermediate, and short time scales. J. Atmos. Sci., 41, 961-979.

Bove, M.C., J.B. Elsner, C.W. Landsea, X. Niu, and J.J. O'Brien, 1998: Effects of El Nino on U.S. Landfalling hurricanes, revisited. Bull. Amer. Meteor. Soc., 79, 2477-2482.

Gershunov, A., and T.P. Barnett, 1998: Interdecadal modulation of ENSO teleconnections. Bull. Amer. Meteor. Soc., 79, 2715-2725.

Glass, F.H., 1998: A climatology of widespread heavy rainfall events across Missouri. Transactions of the Missouri Academy of Sciences, 32.

Gray W.M., 1984: Atlantic seasonal hurricane frequency. Part I: El Nino and 30 mb quasi-biennial oscillation influences. Mon. Wea. Rev., 112, 1649-1668.

Gray, W.M., 1998: Hypothesis on the cause of global multidecadal climate change. Preprints of the Ninth Symposium on Global Change Studies, 271-275.

Hoskins, B.J., I.N. James and G.H. White, 1983: The shape, propagation, and mean-flow interaction of large-scale weather systems. J. Atmos. Sci., 40, 1595-1612.

Houghton, J.T., L.G. Meira-Filho, B.A. Callander, N. Harris, A. Kattenberg and K. Maskell, 1996: Climate change 1995: The science of climatic change, contributions of working group I to the Intergovernmental Panel on Climate Change. Cambridge University Press, 572 pp.

Hu, Q., C.M. Woodruff and S.E. Mudrick, 1998: Intercedal variations of annual precipitation in the central United States. Bull. Amer. Meteor. Soc., 79, 221-229.

Hurrell, J.W., 1995: Decadal trends in the North Atlantic Oscillation: Regional temperatures and precipitation. Science, 269, 676-679.

Hurrell, J.W., 1996: Influence of variations in extratropical wintertime teleconnections on northern hemisphere temperatures. Geophys. Res. Let., 23, 665-668.

Karl, T.R., and co-authors, 1993: A new perspective on recent global warming: Asymmetric trends of daily maximum and minimum temperature. Bull. Amer. Meteor. Soc., 74, 1007-1023.

Karl, T.R., R.W. Knight, D.R. Easterling and R.G. Quayle, 1996: Indicies of climate change for the United States. Bull. Amer. Meteor. Soc., 77, 279-291.

Karl, T.R., and R.W. Knight, 1998: Secular trends of precipitation amount, frequency, and intensity in the United States. Bull. Amer. Meteor. Soc., 79, 231-241.

Kerr, R.A., 1999: Big El Ninos ride the back of slower climate change. Science, 283, 1108-1109.

Key, J.R., and A.C.K. Chan, 1999: Multidecadal global and regional trends in 1000 hPa and 500 hPa cyclone frequencies. Geophys. Res. Lett., 26, 2053 - 2056.

Kung, E.C., C.C. DaCamara, W.E. Baker, J. Susskind, and C.K. Park, 1990: Simulations of winter blocking episodes using observed sea surface temperatures. Quart. J. Roy Meteor Soc., 116, 1053-1070.

Kung, E.C., W. Min, J. Susskind, and C.K. Park, 1992: An analysis of simulated summer blocking episodes. Quart. J. Roy Meteor Soc., 118, 351-363.

Kung, E.C., J. Susskind and C.C. Dacamera, 1993: Prominent northern hemisphere winter blocking episodes and associated anomaly fields of sea surface temperatures. Terr. Atmos. Ocean Sci., 4, 273-291.

Lau, N.C., 1997: Interactions between global SST anomalies and mid-latitude atmospheric circulation. Bull. Amer. Meteor. Soc., 78, 21-33.

Livezey, R.E., M. Masutani, A. Leetma, H. Rui, M. Ji and A. Kumar, 1997: Teleconnective response of the Pacific-North American region atmosphere to large central equatorial Pacific SST anomalies. J. Climate, 10, 1787-1820.

Lupo, A.R. and P.J. Smith, 1995: Planetary and synoptic-scale interactions during the life cycle of a mid-latitude blocking anticyclone over the North Atlantic. Tellus, 47A, 439-456.

Lupo, A.R. and L. Bosart, 1999: An analysis of a relatively rare case of continental blocking. Quart. J. Roy Meteor Soc., 125, 107-138.

McPhaden, M.J., 1999: Climate oscillations: Genesis and evolution of the 1997-1998 El Nino. Science, 283, 950-954.

Metze, D.M., and F.H. Glass, 1998: A climatological study of heavy snowfall events across Missouri. Transactions of the Missouri Academy of Sciences, 32.

Minobe, S., 1997: A 50-70 year climatic oscillation over the North Pacific and North America. Geophys. Res. Lett., 24, 683-686.

Mo, K.C., and R.E. Livezey, 1986: Tropical-extratropical geopotential height teleconnections during the Northern Hemisphere winter. Mon. Wea. Rev., 114, 2488-2515.

Nakamura, H., G. Lin, and T. Yamagata, 1997: Decadal climate in the Northern Pacific during recent decades. Bull. Amer. Meteor. Soc., 78, 2215-2226.

Namais, 1982: Anatomy of great plains protracted heat waves (especially the 1980 U.S. summer summer drought). Mon. Wea. Rev., 110, 824-838.

Neter, J., W. Wasserman, and G.A. Whitmore, 1988: Applied Statistics, 3rd Edition. Boston Allyn and Bacon Press, 1006 pp.

O'Brien, J.J., T.S. Richards and A.C. Davis, 1996: The effect of El Nino on U.S. landfalling hurricanes. Bull. Amer Meteor Soc., 77, 773-774.

O'Brien, J.J., 1998:

Press, W.H., B.P. Flannery, S.A. Teukolsky and W.T. Vetterling, 1988: Numerical recipes in C: The art of scientific computing. Cambridge University Press, New York, N.Y., 735 pp.

Quiroz, R.S., 1987: The association of stratospheric warming with tropospheric blocking. J. Geophys. Res., 91, 5277-5285.

Schmitz, W.J., 1996: On the world ocean circulation: Volume 1, Some global features/North Atlantic circulation. WHOI Report., 96-03, 140 pp.

Singer, F., 1997: Hot talk, cold science. Global warming unfinished debate. The Independent Institute, Oakland, CA, 110 pp.

Wallace, J.M., and D.S. Gutzler, 1981: Teleconnections in the geopotential height field during the northern hemisphere winter. Mon. Wea. Rev., 109, 784-812.

Wunsch, C., 1992: Decade-to-century changes in ocean circulation. Oceanography, 5, 99-106.

Wunsch, C., 1999: The interpretation of short climate records, with comments on the North Atlantic and Southern Oscillations. Bull. Amer. Met. Soc., 80, 245-256.

Zishka, K.M., and P.J. Smith, 1980: The climatology of cyclones and anticyclones over North America and surrounding ocean environs for January and July, 1950 - 1977. Mon. Wea. Rev., 108, 387 - 401.

Footnotes

(1.) The website for the Center for Ocean-Atmospheric Prediction Studies is at: http://masig.fsu.edu].

                     A list of years in this study
                        separated by ENSO phase.
La Nina (LN)  Neutral (NEU)  El Nino (EN)
1949              1950           1951
1954              1952           1957
1955              1953           1963
1956            1958-1962        1965
1964              1966           1969
1967              1968           1972
1970              1974           1976
1971            1977-1981        1982
1973              1983           1986
1975              1984           1987
1988              1985           1991
                  1989           1997
                  1990           1998
                1992-1996
                          Phases of the North
                       Pacific Oscillation (NPO).
NPO PHASE  PERIOD OF RECORD
 Phase 1      1933-1946
 Phase 2      1947-1976
 Phase 1      1977-1998
 Phase 2        1999-
                The total number of seasonal and overall
                snowfall events over Northwest Missouri.
          Fall  Winter  Spring  All
Moderate  22    169     47      238
Heavy     10    84      30      124
Extreme   1     22      6       29
Total     33    275     83      391
                    Snowfalls separated by calendar
                        decade for fall months.
Category    50's     60's     70's     80's      90's
Moderate  5/(0.5)   1/(0.1)  4/(0.4)  7/(0.7)  4/(0.44)
Heavy        0      2/(0.2)  5/(0.5)     0     3/(0.33)
Extreme   1/(0.1)      0        0        0     1/(0.11)
Total     6/(0.6)   3/(0.3)  9/(0.9)  7/(0.7)  8/(0.89)
                       As in Table 4a, except for
                             winter months.
Category    50's      60's      70's      80's      90's
Moderate  43/(4.3)  28/(2.8)  31/(3.1)  37/(3.7)  29/(3.2)
Heavy     12/(1.2)  20/(2.0)  22/(2.2)  17/(1.7)  13/(1.8)
Extreme    6/(0.6)   4/(0.4)   6/(0.6)   2/(0.2)   2/(0.2)
Total     61/(6.1)  52/(5.2)  59/(5.9)  56/(5.6)  44/(4.9)
                       As in Table 4a, except for
                             spring months.
Category    50's      60's      70's      80's      90's
Moderate  10/(1.0)  13/(1.3)  11/(1.1)   7/(0.7)    6/(0.6)
Heavy     11/(1.1)   6/(0.6)   5/(0.5)   5/(0.5)    3/(0.3)
Extreme      0       1/(0.1)   3/(0.3)     0        3/(0.3)
Total     21/(2.1)  20/(2.0)  19/(1.9)  12/(1.2)  12/(1.23)
                 Snowfalls separated by calendar decade
                        for the 50-year sample.
Category    50's      60's      70's      80's      90's
Moderate  58/(5.8)  42/(4.2)  46/(4.6)  51/(5.1)  39/(4.3)
Heavy     23/(2.3)  28/(2.8)  32/(3.2)  22/(2.2)  19/(2.1)
Extreme    7/(.7)    5/(.5)    9/(.9)    2/(.2)    6/(.7)
Total     88/(8.8)  75/(7.5)  87/(8.7)  75/(7.5)  64/(7.1)
                      The total number and average
                     occurrence of snowfalls versus
                         El Nino/La Nina phase.
                 ALL    Moderate   Heavy   Extreme
Cold (LN)       91/8.3   59/5.4    24/2.4   6/0.5
Neutral (NEU)  213.8.2  123/4.7    74/2.8  16/0.6
Warm (EN)       87/7.3   56/4.7    24/2.2   7/0.6
Total          391/8.0  238/4.9   124/2.5  29/0.6
                      Total snowfalls and average
                         occurrence versus ENSO
                         phase for fall months.
                 ALL    Moderate   Heavy   Extreme
Cold (LN)       6/0.54   4/0.36    2/0.18     0
Neutral (NEU)  17/0.68  11/0.44    5/0.20  1/0.04
Warm (EN)      10/0.77   7/0.54    3/0.23     0
Total          33/0.67  22/0.45   10/0.20  1/0.02
                         As in Table 7a, except
                           for winter months.
                 ALL    Moderate   Heavy   Extreme
Cold (LN)       66/6.0   43/3.9   17/1.5    6/0.5
Neutral (NEU)  149/6.0   86/3.4   51/2.0   12/0.5
Warm (EN)       53/4.1   35/2.7   16/1.2   2/0.15
Total          268/5.5  164/3.3   84/1.71  20/0.4
               As in Table 7a, except for spring months.
                 ALL    Moderate   Heavy   Extreme
Cold (LN)      19/1.73  12/1.09    7/0.64     0
Netural (NEU)  47/1.88  26/1.04   18/0.72  3/0.12
Warm (EN)      17/1.31   9/0.69    5/0.38  3/0.23
Total          83/1.69  47/0.96   30/0.61  6/0.12
                 Total number and average occurrence of
               snowfalls stratified by ENSO phase during
               the negative NPO phase (NPO2: 1949-1976).
                 ALL     Moderate  Heavy   Extreme
Cold (LN)       82/8.2    52/5.2   24/2.4   6/0.6
Neutral (NEU)   88/8.0    43/3.9   36/3.3   9/0.8
Warm (EN)       58/8.2    38/5.4   17/2.4   3/0.4
Total          228/8.15  133/4.75  77/2.7  18/0.65
                   As in Table 8 except for positive
                      NPO phase (NPO1: 1977-1997).
                 ALL     Moderate   Heavy   Extreme
Cold (LN)        9/9.0     7/7.0    2/2.0    0/0.0
Neutral (NEU)  125/8.3    80/5.3   38/2.55   7/0.46
Warm (EN)       29/5.8    18/3.6    7/1.4    4/0.8
Total          163/7.75  105/5.0   47/2.25  11/0.56
                 Total number and average occurrence of
                   fall snowfalls for each ENSO phase
                         during each NPO phase.
                 ALL    Moderate  Heavy   Extreme
NPO 2
Cold (LN)       5/0.5    3/0.3    2/0.2      0
Neutral (NEU)   6/0.54   2/0.18   3/0.27   1/0.1
Warm (EN)       5/0.7    4/0.57   1/0.13     0
Total          16/0.57   9/0.32   6/0.21   1/0.04
NPO 1
Cold (LN)       1/1.0    1/1.0      0        0
Neutral (NEU)  12/0.8   10/0.67   2/0.13     0
Warm (EN)       4/0.8    2/0.4    2/0.4      0
Total          17/0.8   13/0.6    4/0.2      0
           As in Table 10, except for winter snowfall events.
                 ALL    Moderate  Heavy   Extreme
NPO 2
Cold (LN)       59/5.9   37/3.7   16/1.6   6/0.6
Neutral (NEU)   56/5.1   27/2.45  22/2.0   7/0.65
Warm (EN)       42/6.0   29/4.15  12/1.7   1/0.15
Total          157/5.6   93/3.3   50/1.8  14/0.5
NPO 1
Cold (LN)        7/7      6/6      1/1       0
Neutral (NEU)   92/6.1   58/3.9   29/1.9   5/0.3
Warm (EN)       19/3.8   12/2.4    4/0.8   3/0.6
Total          118/5.6   76/3.6   34/1.6   8/0.4
           As in Table 10, except for spring snowfall events.
                 ALL    Moderate   Heavy   Extreme
NPO 2
Cold (LN)      18/1.8    12/1.2    6/0.6      0
Neutral (NEU)  26/2.4    14/1.3   11/1.0    1/0.1
Wann (EN)      11/1.57    5/0.7    4/0.6    2/0.28
Total          55/2.0    31/1.1   21/0.75   3/0.1
NPO 1
Cold (LN)       1/1.0      0       1/1.0      0
Neutral (NEU)  21/1.4    12/0.8    7/0.5    2/0.1
Warm (EN)       6/1.2     4/0.8    1/0.2    1/0.2
Total          28/1.33   16/0.8    9/0.4    3/0.15

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]