The Influence of Large-Scale Flow on Fall Precipitation Systems in the Great Lakes Basin

Emily K. Grover Department of Atmospheric, Oceanic, and Space Sciences, University of Michigan, Ann Arbor, Michigan

Search for other papers by Emily K. Grover in
Current site
Google Scholar
PubMed
Close
and
Peter J. Sousounis Department of Atmospheric, Oceanic, and Space Sciences, University of Michigan, Ann Arbor, Michigan

Search for other papers by Peter J. Sousounis in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A synoptic climatology is presented of the precipitation mechanisms that affect the Great Lakes Basin. The focus is on fall because increasing precipitation in this season has contributed to record high lake levels since the 1960s and because the causes can be synoptically evaluated. Precipitation events were identified for the period 1935–95 from NOAA Daily Weather Maps. Precipitation days were classified as one of nine types. Trends in the precipitation classifications, 24-h precipitation totals, and the frequency and intensity of precipitation days and events were analyzed.

It was found that the precipitation increased 15% over the basin and 35% at Grand Rapids, Michigan, from 1935–65 to 1966–95. The increased precipitation was driven by an increase in the amount of precipitation per day (from low pressure systems and warm, stationary, and occluded fronts) and an increase in the frequency of precipitation days (from troughs and cold, warm, stationary, and occluded fronts). All classifications except for isolated convection contributed to the increase. Increases from warm, stationary, and occluded fronts contributed the most.

Analysis of precipitation mechanisms and large-scale circulation features for two 10-yr periods from 1950 to 1959 and from 1980 to 1989 revealed that higher precipitation amounts were associated with a more zonal flow pattern that existed over the United States during 1980–89. This pattern was accompanied by more baroclinicity and moisture over the Rockies, a stronger upper-troposphere subtropical jet, and stronger low-level flow from the Gulf of Mexico. These features allowed a greater number of southern systems with more moisture to influence the region. Specifically, the increased frequency of low pressure systems approaching from the south(west) and their associated more rapid deepening rates allowed more precipitation from warm, stationary, and occluded fronts. The similarities in the synoptic precipitation classifications and precipitation amounts between the two 10-yr periods and the two 30-yr periods examined suggest that more meridional flow was present for much of the 1935–65 period and that more zonal flow was present for much of the 1966–95 period.

Current affiliation: International Research Institute for Climate Research (IRI), Lamont-Doherty Earth Observatory, Columbia University, Palisades, New York

Current affiliation: Weather Services International, Inc., Billerica, Massachusetts

Corresponding author address: Dr. Peter J. Sousounis, Weather Services International, Inc., 900 Technology Park Dr., Bldg. 9, Billerica, MA 01821-4167. Email: psousounis@wsi.com

Abstract

A synoptic climatology is presented of the precipitation mechanisms that affect the Great Lakes Basin. The focus is on fall because increasing precipitation in this season has contributed to record high lake levels since the 1960s and because the causes can be synoptically evaluated. Precipitation events were identified for the period 1935–95 from NOAA Daily Weather Maps. Precipitation days were classified as one of nine types. Trends in the precipitation classifications, 24-h precipitation totals, and the frequency and intensity of precipitation days and events were analyzed.

It was found that the precipitation increased 15% over the basin and 35% at Grand Rapids, Michigan, from 1935–65 to 1966–95. The increased precipitation was driven by an increase in the amount of precipitation per day (from low pressure systems and warm, stationary, and occluded fronts) and an increase in the frequency of precipitation days (from troughs and cold, warm, stationary, and occluded fronts). All classifications except for isolated convection contributed to the increase. Increases from warm, stationary, and occluded fronts contributed the most.

Analysis of precipitation mechanisms and large-scale circulation features for two 10-yr periods from 1950 to 1959 and from 1980 to 1989 revealed that higher precipitation amounts were associated with a more zonal flow pattern that existed over the United States during 1980–89. This pattern was accompanied by more baroclinicity and moisture over the Rockies, a stronger upper-troposphere subtropical jet, and stronger low-level flow from the Gulf of Mexico. These features allowed a greater number of southern systems with more moisture to influence the region. Specifically, the increased frequency of low pressure systems approaching from the south(west) and their associated more rapid deepening rates allowed more precipitation from warm, stationary, and occluded fronts. The similarities in the synoptic precipitation classifications and precipitation amounts between the two 10-yr periods and the two 30-yr periods examined suggest that more meridional flow was present for much of the 1935–65 period and that more zonal flow was present for much of the 1966–95 period.

Current affiliation: International Research Institute for Climate Research (IRI), Lamont-Doherty Earth Observatory, Columbia University, Palisades, New York

Current affiliation: Weather Services International, Inc., Billerica, Massachusetts

Corresponding author address: Dr. Peter J. Sousounis, Weather Services International, Inc., 900 Technology Park Dr., Bldg. 9, Billerica, MA 01821-4167. Email: psousounis@wsi.com

1. Introduction

The Great Lakes play an integral role in many regional aspects, from weather to environmental policy to the economy. One of the most important ways that the Great Lakes govern their region is through their water levels. Record-high lake levels were set in 1973 and 1985 (Quinn 1986; Hitt and Miller 1986). High water levels impact five major areas in the Great Lakes region: shipping, hydropower, recreational boating, shoreline erosion, and the environment (Changnon 1987). Damage reports on the shorelines are much more frequent when lake levels are high (Angel 1995; Meadows et al. 1997). Specifically, the record-high levels set in the 1970s and 1980s, in conjunction with strong storms, caused heavy damage to shorelines, widespread flooding, and destruction of homes (Quinn 1986). These high lake levels and far-reaching impacts have been, in part, a result of an increase in precipitation in the Great Lakes Basin—especially since the mid-1960s (Quinn 1986; Changnon 1987).

Numerous investigators have documented this increasing precipitation regime. Changnon (1987) found a general increasing trend for the Lake Michigan basin beginning at the turn of the century with a sharper increase beginning in the early 1970s. Lettenmaier et al. (1994) identified a significant increase in the annual precipitation of the Great Lakes Basin for the period 1948–88. Karl and Knight (1998) found precipitation increases in the fall season for the Great Lakes region as well as for many other regions in the continuous United States beginning in 1910.

Despite the identification of an increasing precipitation trend in the Great Lakes Basin, little attention has been paid to important details such as its possible correlation with other atmospheric variables. The objectives of this study are to investigate how changes in major precipitation-causing mechanisms (e.g., fronts, troughs, cyclones) have played a role in the increasing precipitation trend of the Great Lakes Basin and to explain these changes by examining the large-scale flow. The time period 1935–95 is examined because it straddles approximately the time (mid-1960s) when precipitation and lake levels began long-term increases. A synoptic classification scheme is implemented to determine the precipitation-forcing mechanisms in the fall months (September–October–November). Section 2 describes the methodology. Section 3 describes the classification results. Section 4 describes how the large-scale flow changed over the period. Section 5 provides a discussion of how changes in the the large-scale flow likely contributed to changes in precipitation characteristics. Section 6 provides a summary and conclusions.

2. Definitions, data, and methodology

For this study, a precipitation day was defined as a day where the 24-h precipitation total was at least a trace at Grand Rapids, Michigan (GRR). Choosing one representative station within a circular Great Lakes region (GLR) that approximated the Great Lakes Basin (GLB; cf. Fig. 1) facilitated the classification of a precipitation day.

Applying the classification scheme to numerous stations within the basin to obtain a more thorough classification climatology was certainly considered. However, it is not unreasonable to assume that the classifications at GRR are representative of the basin because the majority of the possible classifications are synoptic- or meso-α-scale features and would therefore likely be found at other stations in the region (plus or minus a day or two) as it is only 1000 km in diameter. Grand Rapids was chosen because of its long-term availability of precipitation data, its central location within the GLB, and because its precipitation trend is very representative of that for the entire basin. This will be illustrated in more detail in section 3.

A precipitation event was defined as a precipitation day or a group of consecutive precipitation days that were caused by the same surface feature. This definition does not necessarily imply that all of the precipitation days that constitute a precipitation event will have the same classification. For example, a precipitation day may be classified as being caused by a stationary front. If this stationary front becomes a cold front and causes precipitation on the following day, then that day will receive a cold front classification. These two consecutive days would therefore be in the same precipitation event but have different classifications.

The data used in this study was collected from the Daily Weather Maps (DWMs), which are published weekly by the National Oceanic and Atmospheric Association (NOAA 1935–95). The DWMs were chosen because they constituted the only single continuous and hence most internally consistent source of information over the entire period. The type and format of the available information on the DWMs did vary slightly through the 1935–95 investigation period. Generally, the available information of interest included the following: sea level pressure, surface temperature, surface dewpoint, surface wind, cloud cover, weather, surface synoptic features, current areas of precipitation, and 500-hPa heights—all valid at 1200 UTC; previous 24-h precipitation totals, and 24-h high and low temperatures. A slight gap from 1941 to 1943 in the DWMs 24-h precipitation totals required the use of 24-h precipitation totals from the National Climatic Data Center's (NCDC's) Summary-of-the-Day TD-3200 data series from the cooperative climatological network for that period. Another source of data that was especially useful for examining average large-scale conditions was the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis data (Kalnay et al. 1996).

The precipitation day definition was used as a guideline for data collection. For every precipitation day in the fall months from 1935 to 1995, classifications and the 24-h precipitation total from GRR were recorded from the DWMs. The investigation focused on the fall months for several reasons. First, the record lake levels in 1985 were largely a result of extreme precipitation amounts in the fall and winter (Quinn 1986). Second, the historical precipitation trend over the entire region is impressive and distinct for the fall season (cf. Karl and Knight 1998) and accounts for 43% of the annual increase (winter accounts for 10% of the annual increase). Third, the mechanisms that are responsible for precipitation in the fall are generally connected to large-scale dynamical forcing mechanisms, as opposed to small-scale convective forcing mechanisms, and would therefore be easier to identify and be more representative for the entire region.

All of the precipitation days in the period of interest were classified to determine which changes in forcing mechanisms and features are primarily responsible for the increase in fall precipitation in the GLB. The possible precipitation day classifications are cold front, warm front, stationary front, occluded front, low pressure, surface trough, lake effect, isolated convection, and indistinguishable. These classifications span the range of mechanisms that cause precipitation in the GLR. The authors are aware that lake-effect and isolated convection are not generally included in synoptic climatologies, but they were included here because they are important when considering fall precipitation in the GLR. The classifications applied in this study are similar to those used by Shaw (1962), Smithson (1969), and Fiorentino and Havens (1977). Fiorentino and Havens (1977) subjectively evaluated precipitation in New England over a 5-yr period using the same frontal classifications as those utilized in this study and 3 other classifications: warm air mass, cold air mass, and warm tropical (e.g., tropical cyclone).

The flow chart in Fig. 2 describes the logic used to classify the precipitation days. The first step, and possibly the most important one, in classifying the precipitation days, was to compare the 24-h precipitation pattern with the surface features. If the pattern in the vicinity of GRR indicated that a single front as depicted on the DWMs was responsible for the precipitation, then the precipitation day was given the appropriate frontal classification (e.g., cold, warm, stationary, occluded) as determined by the fronts indicated on the DWMs (cf. Figs. 2 and 3a–d).

If multiple fronts contributed to the precipitation on any given day and the fronts were associated with a closed low (e.g., identified by at least one closed isobar at 4-hPa intervals) within or outside the region; or, if the precipitation was not associated with any fronts but was accompanied by a closed low where any portion1 of the closed contours surrounding the low center was in the region for the majority of the 24-h period, then the precipitation day was given a low pressure classification (cf. Figs. 2 and 3e,f).

If the precipitation was not frontal but was associated with a low that was not in the basin; or if the precipitation was not frontal but was associated with a trough in the basin, and lake-effect conditions were not present (see below), then the precipitation day was given a trough classification (cf. Figs. 2 and 3g).

If the precipitation was not frontal, and a closed low was not in the basin, and lake-effect conditions were present, then the precipitation day was given a lake-effect classification (cf. Figs. 2 and 3h). Lake-effect conditions included 1) the precipitation amount at GRR being significantly higher than at surrounding leeward stations (e.g., Detroit, Michigan—DTW, Fort Wayne, Indiana—FWA), and especially those windward of Lake Michigan (e.g., Milwaukee, Wisconsin—MKE; Wausau, Wisconsin—AUW; Chicago, Illinois—ORD) and 2) the GRR surface observations exhibiting a wind direction between 210° and 330° and a surface temperature at least 5°C less than the climatological temperature of Lake Michigan (5°–15°C; cf. Saulesleja 1986).

If the precipitation was isolated and accompanied by conditions favorable for convection, then it received an isolated convection classification (cf. Figs. 2 and 3i). Conditions such as high/low temperatures, dewpoint temperature, and the presence of cyclonic southwesterly flow were considered when determining whether convection was favorable. Conditions at upper levels did not factor into the decision. If the cause of the precipitation could not be identified as either lake-effect or isolated convection, then the precipitation day received an indistinguishable classification (cf. Figs. 2 and 3k).

If multiple fronts were responsible for the precipitation but a closed low was not in the region, then the precipitation day was also given an indistinguishable classification (cf. Figs. 2 and 3j). Using the indistinguishable classification for multiple front cases that were not associated with a closed low avoided having to deal with a constraint in the DWMs. Analyses of the results would be complicated by precipitation days receiving multiple classifications because it raises the question of how much of the 24-h precipitation total was caused by each feature. This question is unanswerable based on the information available from the DWMs. Therefore, this problem is avoided by giving multiple front precipitation days that are not associated with a low, an indistinguishable classification.

There were occasional instances where a combination of a few classifications may have offered the most correct explanation of the cause of the precipitation. For example, if a closed low had recently moved through the region, then the amount of resulting precipitation could have been a result of the low initially and then enhanced by lake-effect conditions. In such a case, the precipitation day would receive a low pressure classification. In this study, the interest lay more in possible large (synoptic, meso α) scale changes rather than smaller (meso β, meso γ) scale changes. Therefore, an event such as this was given the classification that corresponded to the larger scale when two classifications were possible. The exception to this rule was when two different types of fronts were involved, as described above. In that case, the indistinguishable classification was used, since nearly all fronts associated with synoptic-scale lows have essentially the same scale.

The classifications were made subjectively using the above-mentioned criteria in a consistent manner. The subjectivity facilitated the classification of difficult situations (Barry and Perry 1973; Yarnal 1993). There was some concern regarding the consistency with which fronts were analyzed on the DWMs over the 61-yr period because fronts have been drawn manually on the DWMs since their appearance in the DWM series (1941). However, because their placement on DWMs is based on specific criteria that have not changed over the period of interest (Petterssen 1940; National Weather Service 1979; Kocin et al. 1991) this concern was obviated. Prior to 1941, classifications were based on frontal analyses performed by the authors using Petterssen's criteria (1940).

3. Precipitation classifications

The trend in fall precipitation for GRR and for the entire GLB can be seen in Fig. 4. A general increasing trend is present beginning in the mid-1950s with steeper rates of increase in the mid- to late 1960s at GRR and from the mid-1970s to the early 1990s for both GRR and the GLB. There is also a brief decreasing trend in the early 1970s for both plots. The precipitation trends are present in all three months of interest, though they are slightly more prominent in September and November than in October (not shown). The overall regime at GRR is remarkably representative of that for the GLB; there is a strong positive correlation of 0.82 between the two datasets. The strong correlation and the synoptic (scale) nature of the classifications add credibility that variations in the synoptic mechanisms as determined at GRR are relevant for the entire GLR. The average fall precipitation total at GRR was 194 mm for the 1935–65 time period and 262 mm for the 1966–95 time period. This is a 35% increase in seasonal precipitation and is significant to the 5% level based on a z test. The beginning of a decreasing trend in precipitation can be seen in the early 1990s in Fig. 4. This drop in precipitation has contributed to a decline in lake levels as well.

Table 1 illustrates the average seasonal frequency and the precipitation amount per precipitation day (PAPPD) from each classification. A change in the average frequencies was calculated by subtracting the average for the 1935–65 period from the average for the 1966–95 period. Therefore, a positive change indicates an increase in the average seasonal frequency of the classification between the two time periods. The low pressure, lake-effect, and isolated convection classifications had negative changes while the cold front, warm front, stationary front, occluded front, and trough classifications had positive changes. The dominant role of low pressure and cold front classifications is evident as they account for 56% of all classifications over the 61-yr period. However, the dominance of the low pressure classifications decreases as it contributes 29% of all classifications in the 1935–65 period and 23% of all classifications in the 1966–95 period. Note that the 0% precipitation contribution shown for isolated convection classifications in Table 1 is due to the overwhelming number of isolated convection classified days that were trace events (precipitation days that received only trace amounts of precipitation) and offered no measurable precipitation. The combined change in frequencies for all classifications of nearly five precipitation days per season per time period is indicative of an increasing precipitation regime throughout the 61-yr investigation period (not shown) and is significant to the 5% level. Figure 5 illustrates that the changes in the synoptic frequencies of the five most precipitation significant categories in Table 1 are not the result of smooth (linear) trends for all classifications over the 61-yr period. Some classifications exhibit more of a linear trend (warm front, stationary front) than others (low, trough) and no two classifications are highly correlated for the 61-yr period.

The dominance of low pressure and cold front classifications and the relatively low frequency of the other classifications described above is understandable. For example, the relatively low number of warm front classifications exists because a cyclone typically must have moved in a sharp northeastward direction (from southwest of the region) in order for a precipitation day to receive a warm front classification (cf. Fig. 3b). This is uncommon given that the primary cyclone track still lies to the north of the region through October (Reitan 1974). Stationary fronts are also uncommon in the GLR in the fall. In order for a stationary front to develop, geopotential height and thickness (temperature) contours must be parallel to each other. Fall is a transition season in the GLR and is characterized by baroclinic zones that make this alignment of height and thickness lines relatively uncommon. Trough classifications are uncommon in the fall in the GLR because sharp temperature contrasts during that season mean that a trough will likely have a cold front and, therefore, such a case would receive a cold front classification.

The classification frequencies are understandably different from those of Fiorentino and Havens (1977). Differences in methodology and geographical locations explain the different results. For example, they did not consider one of the most important mechanisms in this study, extratropical low pressure systems, as a separate category. Additionally, they found that 60% of all precipitation was warm frontal in winter as a result of the strong land–ocean temperature contrasts.

Table 1 also illustrates the average PAPPD or intensity for each classification. Intensities for most classifications increased: low pressure, warm front, stationary front, occluded front, lake-effect, and indistinguishable classifications had significant positive changes in intensity while cold front, trough, and isolated convection classifications had insignificant positive or negative changes in intensity. These changes are likely due to changes in the duration or the hourly rate of precipitation, which in turn may be due to changes in the strength, speed, or the track of the synoptic feature. Possible explanations for these increases are provided in section 5 although detailed evaluations are beyond the scope of the current study and are left for future research.

The overall precipitation contribution by each of the five most dominant classifications, which represents the combined effects of both changes in frequency and intensity, is illustrated in Fig. 6. Note that while the relative contributions from some classifications increased (warm front and stationary front) and some decreased (low pressure and cold front) between 1935–65 and 1966–95, the absolute contributions from all five classifications increased. The large precipitation contribution by lows and cold fronts is due to their dominance in both frequency and intensity (cf. Table 1). However, while they remain the dominant contributors of precipitation to the seasonal total, their precipitation contribution decreases from 75% of the fall precipitation in 1935–65 to 67% of the fall precipitation in 1966–95. This decrease in relative precipitation contribution by low pressure and cold front mechanisms is coincident with a relative increase from 15% to 23% between the two time periods by warm, stationary, and occluded (occluded not shown in Fig. 6) fronts combined. Seasonal precipitation contributions from these three categories increased from 30 to 63 mm per season between 1935–65 and 1966–95.

The increase from 44.03 precipitation days per season in the 1935–65 period to 48.77 precipitation days per season in the 1966–95 period is also representative of an increase in the number of precipitation events per season from 27.4 events during the 1935–65 time period to 28.6 events during the 1966–95 time period. While this increase in event frequency is not statistically significant, it should be noted that Karl and Knight (1998) demonstrated that an increase in the frequency of precipitation events was responsible for a portion of the increase in annual precipitation. An increase in the average length of precipitation events from 1.60 to 1.70 days (not statistically significant) occurred gradually and also contributed to the increased precipitation over the investigation period. This increase in event length may have come from larger and/or slower weather systems.

An increase in the PAPPD for all mechanisms combined from 4.57 mm day−1 for the period 1935–65 to 5.08 mm day−1 for the period 1966–95 is indicative of the temporal trend shown in Fig. 7. A comparison of Figs. 4 and 7 illustrates a strong relationship (correlation of 0.93) between the PAPPD and the overall precipitation trend for the region, demonstrating that the PAPPD played a very important role in the overall precipitation trend. In order to further investigate this important role, each precipitation day was categorized based on its 24-h precipitation amount. The categories and the number of days in each catagory for the two 30-yr periods are presented in Table 2. While the number of days increased in each catagory from 1935–65 to 1966–95, the fractional increases in the number of category-1, -2, and -3 (light precipitation) days has remained relatively small compared to those for catagory-4 and -5 (heavy precipitation) days.

Figure 8 indicates how each of the precipitation categories contributed to the seasonal precipitation total. The amount of precipitation contribution to the seasonal total by category-4 days starts to increase significantly in the early 1960s. There is a similar trend for category-5 days beginning in the late 1970s. Both of these increases, when compared between the two time periods, are significant to the 5% level. Despite a small difference in the definition of extreme precipitation days, these increases in the contribution toward the seasonal precipitation total by extreme precipitation days are consistent with those found by Karl and Knight (1998). (In this study, category-4 and -5 days are considered to be extreme precipitation days, corresponding to precipitation amounts greater than or equal to 12.70 mm, while Karl and Knight's extreme precipitation days are those falling in the upper 10% of all 24-h precipitation amounts.)

The decrease in the dominance of low pressure and cold front classifications is also exhibited in their contribution to the number of extreme precipitation days. Low pressure and cold fronts composed 78% of all category-4 and -5 days in the 1935–65 time period. This contribution to extreme precipitation days decreased to 66% in the 1966–95 time period. As was the case with total precipitation, the contribution to extreme precipitation days from warm, stationary, and occluded fronts increased the most, leading to a large jump in extreme precipitation days and seasonal precipitation totals.

4. Large-scale flow influences

The influence of the large-scale circulation on regional precipitation has been demonstrated by many studies (e.g., Harman et al. 1980; Rodionov 1994; Mock 1996; Robertson and Ghil 1999). Here too, the higher precipitation regime, as well as some of the changes in precipitation contributions from the different classifications during the 1966–95 period, can be explained in terms of several key differences in the synoptic-scale flow. To this end, it is useful to compare the large-scale conditions for a 10-yr period (1980–89) within the wet 30-yr period (1966–95) to those from a 10-yr period (1950–59) within the dry 30-yr period (1935–65). The 10-yr periods were selected to have enough years to be statistically meaningful but not so many that differences between the wet and dry ones were obscured. Comparison of Tables 3 and 1 shows that the average precipitation characteristics over the 10-yr periods are somewhat representative of those for the 30-yr averages. The correlations for the synoptic classifications and for the precipitation amounts per precipitation day between the 10- and 30-yr periods were 0.49 and 0.51, respectively. The increase of 5.2 precipitation days between the two decades agrees well with the increase of 4.8 days between the two 30-yr periods.

The large-scale comparisons were made using NCEP–NCAR reanalysis data. Figure 9 shows selected large-scale features during the 1980s, and differences for those features relative to the 1950s. A comparison of 500-hPa geopotential heights (Fig. 9b) shows that the 1980s were characterized by a more zonal flow pattern than the 1950s, when 500-hPa heights were 20–25 m higher over the Rockies and 10–15 m lower over the southeastern United States. This result compares favorably with those from Rodionov (1994), who analyzed the influence of large-scale circulation on winter precipitation in the Great Lakes by implementing a modified Pacific–North American (PNA) index. He showed that a negative index, corresponding to zonal flow, was associated with cyclones that originated over the high plains and generated above-normal precipitation. A positive index was associated with cyclones that originated in Alberta, Canada (e.g., Alberta Clippers), and generated below-normal precipitation. In this study, standard monthly PNA indices (Wallace and Gutzler 1981) were averaged over the fall season to obtain values of −0.21 for the wet period (1980s) and 0.16 for the dry period (1950s).

The lower 500-hPa heights over the Rockies during the 1980s were associated with lower temperatures throughout the lower troposphere in that region as well. Temperature differences were as large as −2.5°C at 850 hPa over the southern Rockies (Fig. 9c) and were part of a larger area of lower temperatures that extended southwestward over the eastern Pacific. The lower temperatures and negative height differences translated to increased low-level baroclinicity over the Rockies, which likely contributed to (lee) cyclogenesis.

The increased baroclinicity over the southern Rockies during the 1980s also likely contributed to the stronger winds in the upper troposphere over the Great Lakes Basin. Figure 9a shows that wind speeds at 250 hPa were about 4 m s−1 higher directly over the region than they were during the dry period. These stronger wind speeds were embedded within a more extensive region of higher wind speeds that stretched from the southwestern United States northeastward to New England. The 250-hPa wind speeds that existed during the dry period indicated a weaker jet entrance region in the vicinity of the GLR and one that was also shifted farther north by several hundred kilometers. As a result, the axis of strongest wind speed differences was situated just south of GRR so that most of the region was on the cyclonic side of the jet, which is a favorable location for receiving significant precipitation. Farther to the north and west, evidence for a weaker polar jet exists. Wind speeds were ∼5 m s−1 weaker in the area just south of the Gulf of Alaska during the 1980s than during the 1950s. This pattern suggests that the subtropical jet was stronger and the polar jet was weaker during the wet period than during the dry period, which suggests that storms may have developed farther south and hence been able to draw additional moisture from the Gulf of Mexico rather than be limited to just the moisture from the Pacific.

Other large-scale differences include enhanced low-level moisture in the southwestern United States (not shown). Relative humidity values at 700 hPa were higher by nearly 10% during the 1980s as compared to the 1950s. The moister conditions may have been a result of the stronger subtropical jet. Farther to the north, in Canada, relative humidities were about 10% less. The enhanced moisture, combined with the increased low-level baroclinicity in that region, was likely responsible for the increase in precipitation from all of the mechanisms except isolated convection. This is understandable since isolated convection is a mechanism that is not likely to be affected by enhanced synoptic forcing of large-scale precipitation mechanisms.

Finally, the 1980s were characterized by stronger low-level southerly flow over the region (Fig. 9d). This stronger flow likely contributed to enhanced moisture transport from the Gulf of Mexico and warm advection (cf. Figs. 9c and 9d), both of which contribute to precipitation generation. The surface signature of this stronger southerly flow was in the form of a weak trough located just to the west of the GLR and extended from Texas northeastward to Hudson Bay (not shown). During the wet period, the trough separated a high pressure system over the Rockies to the west from a high pressure system over the western Atlantic to the east. During the 1950s, this trough was even weaker and had a more southwest–northeast flow rather than a south–north flow on the eastern side.

Although it is beyond the scope of this study to examine the specific causes for the different large-scale flow patterns, there is some evidence that these patterns and the precipitation amounts from them during the 1950s and 1980s were influenced by the Pacific Decadal Oscillation (PDO). Published literature (e.g., Mantua et al. 1997) indicates that the PDO index was strongly negative during the 1950s and strongly positive during the 1980s. Given that the effects from PDO are similar to but less intense than those from corresponding El Niño–Southern Oscillation (ENSO) phases (Zhang et al. 1997) and given that fall is drier than normal in the Great Lakes region during the negative ENSO phase and wetter than normal during the positive phase (Climate Prediction Center 2001), it is likely that the lower amounts of fall precipitation in the 1950s and the higher amounts in the 1980s were at least partly the result of the PDO.

5. Discussion

The characteristic differences in the synoptic flow between the 1980s and 1950s help provide a conceptual explanation for understanding how large-scale flow influenced precipitation event frequency and intensity and also the responsible mechanism (e.g., front). The 1980s were characterized by lower geopotential heights over the Rockies that resulted in enhanced low-level baroclinicity, which led to stronger winds near the tropopause. The enhanced winds at upper levels and the enhanced baroclinicity at low levels were both likely more favorable for the development of ascent, clouds, and lee cyclogenesis over the southern Rockies. These more favorable features are reflected in the fact that the 1980s were characterized by a southward shift in the latitude from which cyclones approached the region (cf. Fig. 10). Other features regarding cyclone frequency in Fig. 10 are consistent with the large-scale flow differences between the 1950s and the 1980s. For example, the increased cyclone frequency along the U.S. east coast during the 1950s is consistent with the presence of a more prominent large-scale trough over the eastern half of the United States during that time.

Differences in the large-scale flow pattern(s) between the 1950s and 1980s are also consistent with many of the changes in warm, cold, occluded, and stationary fronts and the respective amounts of precipitation that fell over the GLB. The greater number of cyclones that developed farther south during the 1980s (wet zonal flow) not only resulted in more cyclones going through the region (GRR) so precipitation was classified as “low pressure,” but also resulted in more cyclones tracking just west of the region (cf. Fig. 10)—so precipitation was classified as “frontal” (e.g., instead of “low pressure”). The type of front depended on the relative position of the low with respect to GRR, its track, and intensity.

Warm fronts typically extend (south)eastward from the low center and typically advance north(east)ward. The precipitation can usually extend for several hundred kilometers ahead (e.g., north) of the actual surface warm front. Thus, cyclones tracking from the southwest and passing just west of the region are more likely to influence GRR with a warm front than cyclones tracking from the northwest and passing just east of the region (cf. Fig. 11). Comparison of Figs. 11a and 11b indicates that the mean frontal position the day before the event (e.g., day T-1) during the 1980s was located several hundred kilometers to the south of the corresponding 1950s position—the result of nearly twice as many warm fronts coming from southern cyclones. The southwesterly flow that was associated with these southern cyclones at upper levels had more moisture associated with it so frontal precipitation was also more intense.

A similar explanation holds for cold fronts, which typically extend south(west)ward from a low center and typically advance (south)eastward. Thus, both cyclone types—those tracking from the northwest and passing just east of the region and those tracking from the southwest and passing just west of the region—can influence GRR with cold frontal precipitation. Figures 11c and 11d reflect the fact that even during the 1950s there were cyclones of both types that generated cold front precipitation at GRR—albeit fewer from the south. During the 1980s, there were more of both types. The increase in cold fronts (and precipitation) from southern cyclones is more easily understood than the increase from northern cyclones. One possible explanation for these latter increases is that for developing low pressure systems (e.g., troughs), the warm southerly flow ahead of the low center or trough axis is associated with more moisture when the system is embedded in zonal flow rather than northwesterly flow—even when the system (e.g., low center) is located north of the region. The latent heat release ahead of a developing system then contributes to an enhanced horizontal temperature gradient (e.g., cold front) and hence to increased precipitation. The increase in cold fronts from northern cyclones also explains the increase in troughs and trough precipitation (not shown).

Although changes in the frequencies of occluded and stationary fronts from the 1950s to the 1980s are not statistically significant, differences in the locations and tracks of the cyclones to which these fronts were attached also influenced precipitation intensity. For example, for occluded fronts, Figs. 12a and 12b illustrate how more cyclones approached from the southwest during the 1980s than during the 1950s. These systems were associated with more moisture and hence generated more precipitation. The lack of any significant change in the frequency of occluded fronts along with an increase in the frequency of lows from the southwest suggests that these lows in general may be less mature at the time of approach than those that influence the region from the northwest. Stationary fronts that influenced the Great Lakes during the 1950s and 1980s were usually linked to two cyclones that straddled the region from southwest to northeast. Figures 12c and 12d suggest that during the 1950s, the northern low was the more prominent of the two, and during the 1980s the southern low was the more prominent of the two. The fact that the stationary fronts that influence the region tend to be supported by two low pressure systems on either end (with one usually being the dominant one) may explain the lack of any significant change in frequency.

Figures 11 and 12 illustrate that for all frontal types, high pressure at the surface along the east coast of the United States was more influential (e.g., stronger, closer to the region, and more meridional in shape) during the 1980s, which is consistent with the positive height perturbations at 500 hPa (cf. Fig. 9a) as well as with the reduced storm activity (cf. Fig. 10b) in that region. This feature likely contributed to the stronger southerly flow and to the increased precipitation during the 1980s.

Construction of figures similar to Figs. 11 and 12 but for the entire 61-yr period is tedious but is one way to partially circumvent the nonexistence of upper-air data prior to 1948 in order to deduce some relevant information about what the large-scale flow must have been like and how frontal characteristics were influenced over the entire period of study. A less tedious method may be to combine the results from the two decades that have been evaluated comprehensively with a long-term evaluation of cyclone activity in the Great Lakes Basin. For example, Angel (1996), Angel and Isard (1997, 1998), and Isard et al. (2000) have shown that strong cyclone activity from the 1920s to the 1940s was relatively low in the Great Lakes Basin. Because strong cyclones approach from the southwest, it may be deduced that the large-scale flow during this time was more like that during the 1950s than it was during the 1980s. The fact that the frontal activity during the 1930s–40s was more like that of the 1950s than it was like that of the 1980s is consistent with this deduction. Still, complications (discrepancies) exist between the 10- and 30-yr comparisons, which suggest that additional study is needed to understand how other aspects of the large-scale flow influence fronts and other precipitation mechanisms in the Great Lakes Basin.

6. Summary and conclusions

This study was conducted to understand the cause of the increase in fall precipitation in the Great Lakes Basin since the mid-1960s. It incorporated data from September–November for the period 1935–95 and utilized data from Grand Rapids, Michigan, to represent conditions over the Great Lakes Basin. Precipitation was identified and assessed based on NOAA Daily Weather Maps. Precipitation days were classified according to the synoptic-forcing mechanism that was responsible for the precipitation.

The increasing precipitation trend was documented at both local (GRR) and regional (GLB) levels. Increases of 35% and 15% in fall precipitation totals were exhibited, respectively. All precipitation classifications, with the exception of isolated convection, produced more precipitation in the 1966–95 time period than in the 1935–65 time period. The relative dominance of the low pressure and cold front classifications decreased as warm, stationary, and occluded fronts became more important contributors to fall precipitation. This increase in precipitation production by warm, stationary, and occluded fronts played an important role in the increasing precipitation trend.

Trends in the precipitation day variables were also analyzed. It was found that the frequency of precipitation days as well as the length of precipitation events increased. These increases were accompanied by an increase in the average 24-h precipitation total regardless of classification. The frequency of extreme precipitation days also increased significantly, which was driven by increases from warm, stationary, and occluded fronts.

An evaluation of precipitation and large-scale flow was performed using the NCEP–NCAR reanalysis data. Comparison of wet and dry 10-yr periods within each 30-yr period illustrated that higher precipitation amounts were associated with a more zonal 500-hPa pattern, an enhanced upper-level subtropical jet, an increase in moisture and low-level baroclinicity, and stronger low-level southerly flow. These features contributed to an increase in southern cyclones and hence to an increase in the frequency of and the precipitation from various frontal types. Existing long-term cyclone analyses suggest that the 30-yr periods had large-scale flow characteristics similar to the corresponding 10-yr periods.

It is concluded that significant changes in the large-scale flow between 1935–65 and 1965–95 caused significant changes in the characteristics of precipitation mechanisms, which increased the frequency and intensity with which precipitation fell over the Great Lakes Basin during the fall months. Although this study identified a significant change in the large-scale flow pattern as a cause for the change in precipitation characteristics, it did not address the reasons for the change in the flow pattern and it did not address other possible causes. Additional studies that focus on the other seasons, examine more fully the reasons for changes in the large-scale flow (e.g., from various planetary-scale oscillations), and other causes (e.g., global warming, increased cloud condensation nuclei) in the Great Lakes Basin are needed.

Acknowledgments

The authors would like to thank Dr. Frank Quinn from the Great Lakes Environmental Research Laboratory in Ann Arbor, Michigan, for his motivation to conduct this study. This work was sponsored by the Cooperative Institute of Limnology and Ecosystems Research (CILER) under Cooperative Agreement F000575 from the Great Lakes Environmental Research Laboratory, National Oceanic and Atmospheric Administration, U.S. Department of Commerce. The U.S. Government is authorized to produce and distribute reprints for the governmental purposes notwithstanding any copyright notation that may appear herein.

REFERENCES

  • Angel, J. R., 1995: Large-scale storm damage reports on the U.S. shores of the Great Lakes. J. Great Lakes Res., 3 , 287293.

  • Angel, J. R., . 1996: Cyclone climatology of the Great Lakes. Midwestern Climate Center and Illinois State Water Survey Misc. Publ. 172, Champaign, IL, 122 pp. [Available from Illinois State Water Survey, 2204 Griffith Dr., Champaign, IL 61820-7495.].

    • Search Google Scholar
    • Export Citation
  • Angel, J. R., and S. A. Isard, 1997: An observational study of the influence of the Great Lakes on the speed and intensity of passing cyclones. Mon. Wea. Rev., 125 , 22282237.

    • Search Google Scholar
    • Export Citation
  • Angel, J. R., . 1998: The frequency and intensity of Great Lake cyclones. J. Climate, 11 , 6171.

  • Barry, R. G., and A. H. Perry, 1973: Synoptic Climatology: Methods and Applications. Harper and Row, 555 pp.

  • Changnon, S. A. Jr,, 1987: Climate fluctuations and record-high levels of Lake Michigan. Bull. Amer. Meteor. Soc., 68 , 13941401.

  • Climate Prediction Center, cited 2001: El Niño–Southern Oscillation, and its associated links. [Available online at http://www.cpc.ncep.noaa.gov/products/.].

    • Search Google Scholar
    • Export Citation
  • Douglas, A. V., D. R. Dayan, and J. Namias, 1982: Large-scale changes in North Pacific and North American weather patterns in recent decades. Mon. Wea. Rev., 110 , 18511862.

    • Search Google Scholar
    • Export Citation
  • Fiorentino, D. P., and J. M. Havens, 1977: A synoptic-climatological assessment of precipitation in southeastern New England. Natl. Wea. Dig., 2 , 811.

    • Search Google Scholar
    • Export Citation
  • Harman, J. R., R. Rosen, and W. Corcoran, 1980: Winter cyclones and circulation patterns on the Western Great Lakes. Phys. Geogr., 1 , 2841.

    • Search Google Scholar
    • Export Citation
  • Hitt, K. J., and J. B. Miller, 1986: Great Lakes set record high water levels. Geological Survey, Washington, DC, Water Supply Paper 2300, 35–40.

    • Search Google Scholar
    • Export Citation
  • Isard, S. A., J. R. Angel, and G. T. vanDyke, 2000: Zones of origin for Great Lakes cyclones in North America, 1899–1996. Mon. Wea. Rev., 128 , 474485.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors. 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • 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 , 231241.

    • Search Google Scholar
    • Export Citation
  • Kocin, P. J., D. A. Olson, A. C. Wick, and R. D. Harner, 1991: Surface weather analysis at the National Meteorological Center: Current procedures and future plans. Wea. Forecasting, 6 , 289298.

    • Search Google Scholar
    • Export Citation
  • Lettenmaier, D. P., E. F. Wood, and J. R. Wallis, 1994: Hydro-climatological trends in the continental United States, 1948–88. J. Climate, 7 , 586607.

    • Search Google Scholar
    • Export Citation
  • Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific decadal climate oscillation with impacts on salmon. Bull. Amer. Meteor. Soc., 78 , 10691079.

    • Search Google Scholar
    • Export Citation
  • Meadows, G. A., L. A. Meadows, W. L. Wood, J. M. Hubertz, and M. Perlin, 1997: The relationship between Great Lakes water levels, wave energies, and shoreline damage. Bull. Amer. Meteor. Soc., 78 , 675683.

    • Search Google Scholar
    • Export Citation
  • Mock, C. J., 1996: Climatic controls and spatial variations of precipitation in the western United States. J. Climate, 9 , 11111125.

  • National Oceanic and Atmospheric Administration, 1935–95: Daily Weather Maps—Weekly Series. [Available from NOAA/Climate Prediction Center, Washington, DC 20233.].

    • Search Google Scholar
    • Export Citation
  • National Weather Service, 1979: National Weather Service Forecasting Handbook No.1—Facsimile Products. NOAA NWS FHB 1, 281 pp.

  • Petterssen, S., 1940: Weather Analysis and Forecasting. McGraw-Hill, 503 pp.

  • Quinn, F. H., 1986: Causes and consequences of the record high 1985 Great Lakes water levels. Preprints, Conf. on Climate and Water Management—A Critical Era and Conf. on Human Consequences of 1985's Climate, Asheville, NC, Amer. Meteor. Soc., 281–284.

    • Search Google Scholar
    • Export Citation
  • Reitan, C. H., 1974: Frequency of cyclones and cyclogenesis for North America. Mon. Wea. Rev., 102 , 861868.

  • Robertson, A. W., and M. Ghil, 1999: Large-scale weather regimes and local climate over the western United States. J. Climate, 12 , 17961813.

    • Search Google Scholar
    • Export Citation
  • Rodionov, S. N., 1994: Association between winter precipitation and water level fluctuations in the Great Lakes and atmospheric circulation patterns. J. Climate, 7 , 16931706.

    • Search Google Scholar
    • Export Citation
  • Saulesleja, A., 1986: Atlas Climatologique des Grands Lacs. Ministre de Approvisionnements et Services Canada, 147 pp.

  • Shaw, E. M., 1962: An analysis of the origins of precipitation in northern England, 1956–1960. Quart. J. Roy. Meteor. Soc., 88 , 53947.

    • Search Google Scholar
    • Export Citation
  • Smithson, P. A., 1969: Regional variations in the synoptic origin of rainfall across Scotland. Scot. Geogr. Mag., 85 , 182195.

  • Trenberth, K. E., 2000: Conceptual framework for changes of extremes of the hydrological cycle with climate change. Climatic Change, 42 , 327339.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemispheric winter. Mon. Wea. Rev., 109 , 784812.

    • Search Google Scholar
    • Export Citation
  • Yarnal, B. M., 1993: Synoptic Climatology in Environmental Analysis: A Primer. Belhaven Press, 195 pp.

  • Zhang, Y., J. M. Wallace, and D. S. Battisti, 1997: ENSO-like interdecadal variability: 1900–93. J. Climate, 10 , 10041020.

Fig. 1.
Fig. 1.

The GLB (heavy line) and GLR (dashed circle). Referenced first-order stations are indicated by their three-letter identification code

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 2.
Fig. 2.

Flow chart representing the precipitation day classification methodology. “Circled” letters at the bottom correspond to the panels in Fig. 3

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 3.
Fig. 3.

Schematics of different classifications used for the study: (a) cold front, (b) warm front, (c) stationary front, (d) occluded front, (e) low pressure (centered in region), (f) low pressure (centered outside region), (g) trough, (h) lake effect, (i) isolated convection, (j) indistinguishable (multifront), (k) indistinguishable (does not fit other classifications). Conventional surface features are shown and are valid at the end of the 24-h period to be classified. Hatched areas indicate areas of active precipitation. Arrows indicate path of surface cyclone.

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 4.
Fig. 4.

The 7-yr running average of fall precipitation totals for GRR and the GLB

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 5.
Fig. 5.

The 7-yr running average of seasonal frequency of precipitation classifications. Only the five most common classifications are included

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 6.
Fig. 6.

The 7-yr running average of precipitation contributed by precipitation classifications to seasonal precipitation total. Only the five highest contributing classifications are included

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 7.
Fig. 7.

The 7-yr running average of average amount of precipitation per precipitation day at GRR

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 8.
Fig. 8.

The 7-yr running average of precipitation amount contributed by precipitation categories to seasonal precipitation total

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 9.
Fig. 9.

Average conditions for 1980–89 fall seasons (heavy contours) and differences between 1980–89 and 1950–59 fall seasons (thin contours) for (a) 250-hPa wind speeds (heavy contours every 2.5 m s−1 and thin contours every 1 m s−1), (b) 500-hPa geopotential heights (heavy contours every 6 dam and thin contours every 5 m); (c) 850-hPa temperatures (heavy contours every 4°C and thin contours every 0.5°C); (d) 850-hPa υ-component winds (heavy contours every 1 m s−1 and thin contours every 0.5 m s−1). Shaded regions indicate significant positive [in (a) and (d)] or negative [in (b) and (c)] differences. [Provided by the NOAA–CIRES Climate Diagnostics Center, Boulder, Colorado, at their Web site at http://www.cdc.noaa.gov/]

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 10.
Fig. 10.

Number of cyclones per 2.5° × 2.5° box found from NCEP–NCAR reanalysis daily data during Sep–Nov for (a) 1950–59 and (b) 1980–89

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 11.
Fig. 11.

Average sea level pressure (solid—hPa), surface temperature (dashed—°C), and daily precipitation (see legend for shading definitions) for (a) warm front cases during the 1950s, (b) warm front cases during the 1980s, (c) cold front cases during the 1950s, and (d) cold front cases during the 1980s at day T-1 from NCEP–NCAR reanalyses. Small circles indicate locations of corresponding low centers. Small stars indicate locations of significant high centers associated with the fronts

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Fig. 12.
Fig. 12.

Average sea level pressure (solid—hPa), surface temperature (dashed—°C), and daily precipitation (see legend for shading definitions) for (a) occluded front cases during the 1950s, (b) occluded front cases during the 1980s, (c) stationary front cases during the 1950s, and (d) stationary front cases during the 1980s at day T-1 from NCEP–NCAR reanalyses. Small circles indicate locations of corresponding low centers

Citation: Journal of Climate 15, 14; 10.1175/1520-0442(2002)015<1943:TIOLSF>2.0.CO;2

Table 1. 

Average precipitation day (PD) frequency and PAPPD for mechanisms per fall season

Table 1. 
Table 2. 

Average number of precipitation days falling in the following 24-h precipitation total categories per fall season: cat 1 (0.25–1.27 mm), cat 2 (1.52–6.10 mm), cat 3 (6.35–12.45 mm), cat 4 (12.4–25.15 mm), cat 5 (≥25.40 mm). Precipitation totals of trace are not included. Therefore, the sum of each time period's number of precipitation category days is less than the average number of total precipitation days for that period

Table 2. 
Table 3. 

Similar to Table 1 but for two 10-yr periods

Table 3. 

1

Because the DWMs were only available every 24 h, the positions of the contours were determined based on interpolation between maps with assistance from the 6-hourly positions of the low centers that were available on most maps.

Save
  • Angel, J. R., 1995: Large-scale storm damage reports on the U.S. shores of the Great Lakes. J. Great Lakes Res., 3 , 287293.

  • Angel, J. R., . 1996: Cyclone climatology of the Great Lakes. Midwestern Climate Center and Illinois State Water Survey Misc. Publ. 172, Champaign, IL, 122 pp. [Available from Illinois State Water Survey, 2204 Griffith Dr., Champaign, IL 61820-7495.].

    • Search Google Scholar
    • Export Citation
  • Angel, J. R., and S. A. Isard, 1997: An observational study of the influence of the Great Lakes on the speed and intensity of passing cyclones. Mon. Wea. Rev., 125 , 22282237.

    • Search Google Scholar
    • Export Citation
  • Angel, J. R., . 1998: The frequency and intensity of Great Lake cyclones. J. Climate, 11 , 6171.

  • Barry, R. G., and A. H. Perry, 1973: Synoptic Climatology: Methods and Applications. Harper and Row, 555 pp.

  • Changnon, S. A. Jr,, 1987: Climate fluctuations and record-high levels of Lake Michigan. Bull. Amer. Meteor. Soc., 68 , 13941401.

  • Climate Prediction Center, cited 2001: El Niño–Southern Oscillation, and its associated links. [Available online at http://www.cpc.ncep.noaa.gov/products/.].

    • Search Google Scholar
    • Export Citation
  • Douglas, A. V., D. R. Dayan, and J. Namias, 1982: Large-scale changes in North Pacific and North American weather patterns in recent decades. Mon. Wea. Rev., 110 , 18511862.

    • Search Google Scholar
    • Export Citation
  • Fiorentino, D. P., and J. M. Havens, 1977: A synoptic-climatological assessment of precipitation in southeastern New England. Natl. Wea. Dig., 2 , 811.

    • Search Google Scholar
    • Export Citation
  • Harman, J. R., R. Rosen, and W. Corcoran, 1980: Winter cyclones and circulation patterns on the Western Great Lakes. Phys. Geogr., 1 , 2841.

    • Search Google Scholar
    • Export Citation
  • Hitt, K. J., and J. B. Miller, 1986: Great Lakes set record high water levels. Geological Survey, Washington, DC, Water Supply Paper 2300, 35–40.

    • Search Google Scholar
    • Export Citation
  • Isard, S. A., J. R. Angel, and G. T. vanDyke, 2000: Zones of origin for Great Lakes cyclones in North America, 1899–1996. Mon. Wea. Rev., 128 , 474485.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors. 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • 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 , 231241.

    • Search Google Scholar
    • Export Citation
  • Kocin, P. J., D. A. Olson, A. C. Wick, and R. D. Harner, 1991: Surface weather analysis at the National Meteorological Center: Current procedures and future plans. Wea. Forecasting, 6 , 289298.

    • Search Google Scholar
    • Export Citation
  • Lettenmaier, D. P., E. F. Wood, and J. R. Wallis, 1994: Hydro-climatological trends in the continental United States, 1948–88. J. Climate, 7 , 586607.

    • Search Google Scholar
    • Export Citation
  • Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific decadal climate oscillation with impacts on salmon. Bull. Amer. Meteor. Soc., 78 , 10691079.

    • Search Google Scholar
    • Export Citation
  • Meadows, G. A., L. A. Meadows, W. L. Wood, J. M. Hubertz, and M. Perlin, 1997: The relationship between Great Lakes water levels, wave energies, and shoreline damage. Bull. Amer. Meteor. Soc., 78 , 675683.

    • Search Google Scholar
    • Export Citation
  • Mock, C. J., 1996: Climatic controls and spatial variations of precipitation in the western United States. J. Climate, 9 , 11111125.

  • National Oceanic and Atmospheric Administration, 1935–95: Daily Weather Maps—Weekly Series. [Available from NOAA/Climate Prediction Center, Washington, DC 20233.].

    • Search Google Scholar
    • Export Citation
  • National Weather Service, 1979: National Weather Service Forecasting Handbook No.1—Facsimile Products. NOAA NWS FHB 1, 281 pp.

  • Petterssen, S., 1940: Weather Analysis and Forecasting. McGraw-Hill, 503 pp.

  • Quinn, F. H., 1986: Causes and consequences of the record high 1985 Great Lakes water levels. Preprints, Conf. on Climate and Water Management—A Critical Era and Conf. on Human Consequences of 1985's Climate, Asheville, NC, Amer. Meteor. Soc., 281–284.

    • Search Google Scholar
    • Export Citation
  • Reitan, C. H., 1974: Frequency of cyclones and cyclogenesis for North America. Mon. Wea. Rev., 102 , 861868.

  • Robertson, A. W., and M. Ghil, 1999: Large-scale weather regimes and local climate over the western United States. J. Climate, 12 , 17961813.

    • Search Google Scholar
    • Export Citation
  • Rodionov, S. N., 1994: Association between winter precipitation and water level fluctuations in the Great Lakes and atmospheric circulation patterns. J. Climate, 7 , 16931706.

    • Search Google Scholar
    • Export Citation
  • Saulesleja, A., 1986: Atlas Climatologique des Grands Lacs. Ministre de Approvisionnements et Services Canada, 147 pp.

  • Shaw, E. M., 1962: An analysis of the origins of precipitation in northern England, 1956–1960. Quart. J. Roy. Meteor. Soc., 88 , 53947.

    • Search Google Scholar
    • Export Citation
  • Smithson, P. A., 1969: Regional variations in the synoptic origin of rainfall across Scotland. Scot. Geogr. Mag., 85 , 182195.

  • Trenberth, K. E., 2000: Conceptual framework for changes of extremes of the hydrological cycle with climate change. Climatic Change, 42 , 327339.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemispheric winter. Mon. Wea. Rev., 109 , 784812.

    • Search Google Scholar
    • Export Citation
  • Yarnal, B. M., 1993: Synoptic Climatology in Environmental Analysis: A Primer. Belhaven Press, 195 pp.

  • Zhang, Y., J. M. Wallace, and D. S. Battisti, 1997: ENSO-like interdecadal variability: 1900–93. J. Climate, 10 , 10041020.

  • Fig. 1.

    The GLB (heavy line) and GLR (dashed circle). Referenced first-order stations are indicated by their three-letter identification code

  • Fig. 2.

    Flow chart representing the precipitation day classification methodology. “Circled” letters at the bottom correspond to the panels in Fig. 3

  • Fig. 3.

    Schematics of different classifications used for the study: (a) cold front, (b) warm front, (c) stationary front, (d) occluded front, (e) low pressure (centered in region), (f) low pressure (centered outside region), (g) trough, (h) lake effect, (i) isolated convection, (j) indistinguishable (multifront), (k) indistinguishable (does not fit other classifications). Conventional surface features are shown and are valid at the end of the 24-h period to be classified. Hatched areas indicate areas of active precipitation. Arrows indicate path of surface cyclone.

  • Fig. 4.

    The 7-yr running average of fall precipitation totals for GRR and the GLB

  • Fig. 5.

    The 7-yr running average of seasonal frequency of precipitation classifications. Only the five most common classifications are included

  • Fig. 6.

    The 7-yr running average of precipitation contributed by precipitation classifications to seasonal precipitation total. Only the five highest contributing classifications are included

  • Fig. 7.

    The 7-yr running average of average amount of precipitation per precipitation day at GRR

  • Fig. 8.

    The 7-yr running average of precipitation amount contributed by precipitation categories to seasonal precipitation total

  • Fig. 9.

    Average conditions for 1980–89 fall seasons (heavy contours) and differences between 1980–89 and 1950–59 fall seasons (thin contours) for (a) 250-hPa wind speeds (heavy contours every 2.5 m s−1 and thin contours every 1 m s−1), (b) 500-hPa geopotential heights (heavy contours every 6 dam and thin contours every 5 m); (c) 850-hPa temperatures (heavy contours every 4°C and thin contours every 0.5°C); (d) 850-hPa υ-component winds (heavy contours every 1 m s−1 and thin contours every 0.5 m s−1). Shaded regions indicate significant positive [in (a) and (d)] or negative [in (b) and (c)] differences. [Provided by the NOAA–CIRES Climate Diagnostics Center, Boulder, Colorado, at their Web site at http://www.cdc.noaa.gov/]

  • Fig. 10.

    Number of cyclones per 2.5° × 2.5° box found from NCEP–NCAR reanalysis daily data during Sep–Nov for (a) 1950–59 and (b) 1980–89

  • Fig. 11.

    Average sea level pressure (solid—hPa), surface temperature (dashed—°C), and daily precipitation (see legend for shading definitions) for (a) warm front cases during the 1950s, (b) warm front cases during the 1980s, (c) cold front cases during the 1950s, and (d) cold front cases during the 1980s at day T-1 from NCEP–NCAR reanalyses. Small circles indicate locations of corresponding low centers. Small stars indicate locations of significant high centers associated with the fronts

  • Fig. 12.

    Average sea level pressure (solid—hPa), surface temperature (dashed—°C), and daily precipitation (see legend for shading definitions) for (a) occluded front cases during the 1950s, (b) occluded front cases during the 1980s, (c) stationary front cases during the 1950s, and (d) stationary front cases during the 1980s at day T-1 from NCEP–NCAR reanalyses. Small circles indicate locations of corresponding low centers

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 320 78 11
PDF Downloads 148 54 2