Brian F. O'Hara
Julie L. Adolphson
Tom Reaugh
Ed Holicky



National Weather Service Office
North Webster, Indiana
(Northern Indiana)




This paper is a compilation of nearly five decades of meteorological data which includes information about severe weather phenomena. Tornadoes, hail, high winds, lightning, flooding, and severe winter-weather cases are included. Two sources of information were used to compile the severe weather information. First, the National Weather Service's (NWS) Storm Prediction Center Database (1950 to 1997) was used to compute frequency, time of year and day, and many other types of information regarding severe weather. Second, the National Climatic Data Center's Storm Data, a monthly publication, was also used to gather additional data (NOAA 1959-1997). A plotting program with an associated database was used to supplement the data (Hart 1993). The information in this study is intended to provide guidance to local forecasters (especially newly hired employees), emergency managers, media representatives, and others by illuminating the most frequent occurrences (both spatially and temporally) of severe weather in the Northern Indiana County Warning Area (CWA).

With the NWS Modernization and Associated Restructuring (MAR), many field offices have been consolidated. Weather Service Offices (WSO) and Weather Service Forecast Offices (WSFO) have been combined in order to provide a wider range of services to correspondingly smaller service areas. A forecast office will provide all forecast and warning's services to smaller areas. Each new forecast office will be associated with its own doppler radar. This radar is officially known as Weather Surveillance Radar - 1988 Doppler (WSR-88D) and are popularly referred to as NEXRADs (Next Generation Weather Radars).

Developed during the 1980s, and deployed during the last ten years, these radars offer the latest technology to allow meteorologists to detect areas of precipitation and thunderstorms. The doppler aspect of the radar allows forecasters to see circulations inside of thunderstorms, thereby showing a thunderstorm possibly nearing severe criteria or containing a rotation that could eventually reach the ground as a tornado. This should help increase lead times in the issuance of severe thunderstorm and tornado warnings. The radar's computer processing also allows for the estimation of rainfall accumulations for various time periods. This allows hydrologists and meteorologists to issue flood and flash flood warnings by correlating these rainfall estimates with actual rain gauge readings. Computer algorithms are also being developed to help in the estimation of snowfall.

In 1997, a new WSR-88D radar was installed at the new WSO on State Road 13, two miles north of North Webster, Indiana, in order to better serve the northern Indiana, southern Michigan, and northwest Ohio regions. Specific information on the area served by the new office is included in the next section.

Prior to assuming operational responsibility, the staff of the new office trained over 1200 spotters to identify and report severe weather. With the new radar and the aid of these spotters, it is hoped that the severe weather database will provide continuing accurate severe weather information.


A. Population

The new NEXRAD Weather Service Office (NWSO) in Northern Indiana was created by combining the WSOs located at Ft. Wayne and South Bend, Indiana. NWSO Northern Indiana's CWA is composed of 37 counties in three states. Twenty-four counties are in northern and northeastern Indiana, five are in southwestern Michigan, and eight are in northwestern Ohio (Figure 1). Land use ranges from rural farmland and orchard areas to urban areas. According to the 1990 U.S. Census, six counties in the CWA had populations of over 100,000. The most populous county is Allen in Indiana with a population of 300,836and the least populated county is Pulaski in Indiana with a population of 12,643, both in Indiana. The total population for the CWA is 2,159,634 (Figure 2).

The largest city is Ft. Wayne, Indiana (Allen County) that had a 1990 population of 173,072. The second largest city in the CWA is South Bend (St. Joseph County) with a population of 105,511. Five other cities in the CWA (Lima, Ohio; Elkhart, Mishawaka, Michigan City, and Marion in Indiana) had populations of over 30,000.

The CWA encompasses an area of 15,875 square miles (sq. mi.). The size of the individual counties ranges from 657 sq. mi. (Allen in Indiana) to 165 sq. mi. (Blackford in Indiana) (Figure 3).

Population density varies greatly across the CWA. The most densely populated counties, as can be expected, contain the larger cities. St. Joseph County (Indiana) has the highest population density (540.60 persons per square mile). The second highest is Allen County (Indiana) with a density of 457.89. Rounding out the top five are Elkhart County (Indiana) (336.63), Berrien County (Michigan) (282.62), and Allen County (Ohio) (271.67) (Figure 4). The least populated county, Pulaski also has the lowest population density (29.13). White County (Indiana) (another relatively large rural county) has the second lowest density (46.07) and Paulding County (Ohio) is third with 49.25.

B. Physiography and Land Use

The CWA ranges in elevation from around 580 feet above sea level on the shores of Lake Michigan to around 1200 feet above sea level in Hillsdale County (Michigan). The CWA is bisected from west to southeast by an interesting physiographic feature. The Continental Divide separates two main drainage basins across the southern Great Lakes region. To the north of the divide, rivers drain into the Great Lakes and ultimately into the Atlantic Ocean through the St. Lawrence River. South of the divide, rivers feed into the Ohio and Mississippi Rivers and then into the Gulf of Mexico.

The shoreline of southeastern Lake Michigan consists mainly of sand. In fact, large sand dunes can be found all along eastern Lake Michigan from northwestern Indiana to the northwestern part of Michigan's Lower Peninsula. Lake Michigan is divided into two basins. The deeper of the two is in the northern half of the lake. The shallower basin in the southern third of the lake extends to depths of greater than 500 feet. The water circulation in Lake Michigan consists of two main gyres that seem to be associated with each basin. Water in each gyre circulates in a counterclockwise direction. This results in a northerly longshore current along the southeastern shore of Lake Michigan (Hough 1958). The lake plays an important meteorological role in counties near the lake. More details about this will follow in subsequent sections.

Most of the non-urban areas are devoted to agriculture. Corn and soybeans comprise a large portion of the crop acreage. Wheat is also grown throughout the region. The proximity of Lake Michigan helps to extend the growing season, especially in the northwestern part of the CWA. Southwestern Michigan and northwestern Indiana are known for their large areas devoted to orchards and vineyards. During mild autumns the growing season can last through November.


Severe weather can occur throughout the year across the CWA but typically follows an annual pattern (Figure 5). Thunderstorms and tornadoes have occurred in all months of the year. However, severe thunderstorms and stronger tornadoes can be expected mainly during the spring and summer months. A severe thunderstorm is defined by the NWS as a "thunderstorm that produces a tornado, winds of at least 50 knots (58 mph), and/or hail at least 0.75 inch in diameter" (DOC 1995).

Even though severe weather can occur during any time of the year, certain conditions are necessary for the initiation of strong convection. Almost all severe local storm events are associated with deep convection (Johns and Doswell 1992). Moisture is necessary in order to provide the latent heat to support deep convection. The atmosphere also has to be unstable enough to allow any upward motion in a thunderstorm to continue. Finally, a "trigger" is necessary to start a portion of the lower atmosphere on its journey upward. In the southern Great Lakes region, these conditions are most often found during the spring and summer months.

Derechos, straight-line wind events, typically occur during the warmest part of the year during the summer. A derecho can cause a continuous path of destruction for hundreds of miles and the entire event can last for nearly 24 hours.

Flooding can be a problem at any time of year, but is most common during the spring when rainfall can combine with melting snowcover to produce particularly devastating floods. Flash floods can be extremely destructive and life-threatening due to the quickness with which some rivers rise.

Winter weather can adversely affect people across the southern Great Lakes region. Heavy snow can occur during lake effect events as cold air moves across the relatively warmer Great Lakes during the autumn and early spring when the lakes may not be completely ice-covered. During the depths of winter, arctic outbreaks can bring extremely cold air into the region with temperatures of below zero degrees Fahrenheit (0°F) lasting sometimes for days at a time. Blizzards can paralyze the region for days, making travel on snow-covered roads impossible and cutting off electricity (and heat) to thousands of residents. Each of the above extreme weather types is discussed in further detail in the following sections.


Northern Indiana, northwest Ohio, and extreme southern lower Michigan are especially prone to high wind events. Throughout the last 43 years (1955-1997) there has been a total of 1406 wind events reported in our 37 CWA.

Frequently, wind events are associated with severe thunderstorms. These storms may be individual cells that last only an average of 30 minutes. Other systems that produce damaging winds include lines of storms. These lines can be hundreds of miles long and traverse many miles across the Great Lakes areas. Many of these lines take on a "bow" shape and are referred to as "bow echos." Winds from the rear of the line penetrate to the surface mainly near the apex of the bow, causing extensive damage. Sometimes, circulations develop on the ends of the lines, and are called "book-end vortices." These circulation regions can sometimes spawn weak tornadoes.

Fast moving lines that cover several states are sometimes referred to as derechos (Johns and Hirt 1987). Wind speeds in excess of 60 mph are transported from around 10,000 to 15,000 feet above the ground down to the surface.

About 10% of wind events are a result of high surface winds in association with strong low pressure systems. In order to achieve a balance of forces in the fluid that we call the atmosphere, wind must flow counterclockwise around a low pressure system. If the low pressure is relatively intense, a strong gradient is established and high winds result. These systems can happen at any time of the year, but are most frequent in late fall and early winter.

All wind events were tallied for the period 1955-1997. Overall, more events were reported in June and July (Figure 6). The proximity in the summer of the 500 millibar jet axis near the Great Lakes may contribute to the high number of northwest flow events across this region (Johns 1984). The majority of these events were associated with severe thunderstorms and occurred in the afternoon and evening hours (Figure 7). Interestingly, a large number of the events occurred in the late evening hours after dark. The phenomena most likely to contribute to this number are called nocturnal mesoscale convective systems (MCS). The nighttime cooling of the cloud top's aids in the destabilization needed for nocturnal storms to occur. Other processes involved include the presence of a boundary such as a front or an outflow from another storm. These storms are slow moving and also produce copious rainfall, in addition to strong outflow winds.

As expected, the number of wind events noted in this period is directly correlated to the population (Figure 8 ). Also, the number of reports per year shows an increase with time (Figure 9). The likely reason for this is an increase in population coupled with improved communications with time.


When there is ample moisture in the atmosphere, and especially when there is enough instability, hail can form in thunderstorms and grow to great size. Hail forms when a condensation nucleus passes through supercooled water in a cloud. Hailstones grow in size when they pass through cloud layers of varying water content. A hailstone is kept aloft by the updraft in a cumulonimbus (thunderstorm) cloud. The stronger the updraft, the longer a hailstone will stay suspended in the cloud and continue to collect a coating of water.

As the updraft lifts the hailstone above the freezing level in the cloud, the surface of the hailstone freezes. Conversely, when it falls below the freezing level, the surface melts. Subsequent passes of the hailstone above and below the freezing level, produce the layered pattern associated with hailstones. The hailstone finally falls to the ground when it becomes too large for the thunderstorm updraft to support it.

Once the hailstone falls below the base of the cloud it stops collecting water on its surface. The warmer and drier air below the cloud starts to melt the surface of the hailstone. For hailstones to reach the ground they must be large enough so that they do not melt completely during their descent, therefore to begin with it must be relatively large. This is why large hail is usually experienced with more intense thunderstorms. The updrafts must be relatively strong in these thunderstorms to keep the hailstones aloft long enough to collect many layers of ice and grow to enormous sizes.

Hail with the diameter of golf balls and occasionally (although rarely) with the diameter of softballs can form in some of the more intense thunderstorms. Hail most likely forms and falls to the ground in most of the stronger thunderstorms that are experienced across the Great Lakes region. However, many of these hail occurrences may go unreported simply because no one sees the hail before it melts. This can be especially true in some of the more sparsely populated parts of the CWA.

From 1950 through 1997, 626 hail events were reported in the CWA. This number includes only those reports for which specific sizes were reported. Hail reports have increased throughout the years (Figure 10). In the 1950s and 1960s relatively few hail reports were received by the NWS. This isn't because there were necessarily fewer severe thunderstorms. It is more likely because the NWS did not have as comprehensive a warning verification system in place. Also the NWS may not have conducted post-storm surveys as often or completely as in more recent years. The number of hail reports increases in the 1970s and especially in the 1980s and continues to the present. There is no discernable pattern to the hail events through the years. Some years in the 1950s had no reports of hail (due probably to the reasons just mentioned) and some years in the 1980s and 1990s had over 30 reports of hail occurring throughout the year.

Reports of hail occurrences in Northern Indiana's CWA from 1950 through 1997 follow a fairly typical annual pattern (Figure 11). Hail reports start to increase in March, reaching a maximum in June. This is true for all hail sizes up to 2.75 inches in diameter (Figures 12 and 13). From 1950 through 1997 there have only been 15 reports of hail greater than or equal to 2.75 inches in diameter (Figure 14). Obviously, this is too small of a sample to make a definitive statement on hail frequency but this small sample shows an increase in reports from March to April and then a decrease after June.

Hail reports also showed a typical diurnal pattern (Figure 15). The number of reports increases during the afternoon hours, reaching a peak frequency of occurrence at 4:00 p.m. EST. Relatively high frequencies are also seen at 6:00 p.m. and 7:00 p.m. EST with a sharp decrease throughout the later evening hours. Interestingly, there is a decreased frequency at 5:00 p.m. EST. This is a fairly significant difference, being a decrease of almost 1/3 from the 4:00 p.m. total.

Hail is more frequently reported in the more densely populated parts of the CWA. Allen and St. Joseph Counties in Indiana and Berrien County in Michigan have recorded the most hail occurrences since 1950. Huntington County (Indiana) has also reported many instances of hail and this may be partly due to its relatively large geographic area (Figure 16).


A. Introduction

Tornadoes are one of the most visually impressive weather phenomena known. The small area covered by a tornado's path can experience incredible destruction. Tornadoes are generally associated with supercell storms. Supercells are large, rotating storm systems that can produce a wide variety of severe weather. Many tornadoes in the southern Great Lakes region, however, form in a non-supercell environment. These systems can sometimes be difficult to detect with their relatively smaller mesocyclones (small scale circulations within the storm). New operational and research observing systems will certainly aid meteorologists' ability to understand and predict non-supercell tornadoes (Doswell and Burgess 1993). Specifically, with the introduction of the doppler radars and denser observational networks, the ability to detect tornadoes and non-tornadic mesocyclones should continue to improve. However, a tornado does not have to be preceded by a detectable mid-level mesocyclone. Although, those that are tend to be larger and more intense (Wakimoto and Wilson 1989).

From 1950 to 1997, 408 tornadoes were reported within the modern-day CWA of NWSO Northern Indiana. This is an average of 8½ tornadoes per year. These tornadoes have ranged in strength from numerous small F0 twisters, to a massive funnel that produced F5 destruction1 . The following table contains a description of the F-Scale, which was created by Dr. Ted Fujita, of the University of Chicago:



Wind speed (mph)

F0 <73
F1 73 - 112
F2 113 - 157
F3 158 - 206
F4 207 - 260
F5 >261



1The F5 tornado referred to here is the tornado that severely damaged Dunlap, Indiana on the southeast side of Elkhart on April 11, 1965. This tornado is officially listed in NWS records as an F4. However, in Significant Tornadoes, 1680-1991 it was rated as an F5 (Grazulis 1993). It is the only tornado Grazulis rated as an F5 in the entire CWA over the period of record back to 1834. Because of the extreme devastation caused by this infamous tornado, we have decided to apply Grazulis' rating of F5 to this tornado. All other F-Scale rankings in this paper are those assigned by NWS personnel.

Tornadoes have occurred in every month, during almost every hour of the day, and in every county in the CWA. There have been occasions where as many as 31 tornadoes were reported in a single day.

There have been several tornadoes that occurred in atypical months or at atypical times of the day. On February 18, 1992, an F4 tornado swept away houses near Ohio City, Ohio. Four days before Christmas in 1967 four tornadoes roared across the southeast part of the CWA, causing F3 damage in Grant and Blackford Counties and on January 3, 1950, an F1 tornado touched down in Van Wert County. Elkhart and Grant Counties have both reported F2 strength tornadoes around 6:00 a.m. EST.

The Super Outbreak of April 3, 1974 produced 17 tornadoes and the Palm Sunday Outbreak of April 11, 1965 recorded 14. These two incredible days were significant outliers in the data and created unique problems in getting accurate and helpful information. After noting theses exceptional tornado cases, there are still the majority of tornadoes that fall within more typical temporal boundaries. The following sections will address the climatology of tornadoes to better anticipate their behavior with respect to the Northern Indiana CWA.

B. Tornadoes per Year

Figure 17 shows the total number of tornadoes reported for each year in KIWX's CWA. The most tornadoes occurred in the year of the Palm Sunday Outbreak (1965) and the year of the Super Outbreak (1974). In 1965, 31 tornadoes were recorded, with the Palm Sunday Outbreak making up14 of them (45%). In 1974, 17 out of the 23 tornadoes that year (an impressive 74%) occurred during the Super Outbreak.

The third most notable tornado year was 1992 with 22 tornadoes. The tornadoes in 1992 were fairly evenly spread throughout the year. The first occurred in February with the last in November. On July 12 and again October 8, there were four tornadoes that were the most on one day during the year.

No tornadoes are on record for the year 1952. It is highly improbable that the entire CWA was tornado-free for that entire year. The reasons for this anomalous year are likely similar to those mentioned in previous sections. Tornadoes occurred in sparsely populated areas and reporting practices were less developed in the 1950s.

On June 8, 1953, a tornado tore through Flint, Michigan and became the last U.S. single tornado to kill more than 100 people (115). This event combined with 13 other reported tornadoes in 1953, for a total of 14 tornadoes that year. The number of annual tornadoes dropped below five only two more times (in 1955 and 1972) until the early 1980s.

This figure shows the years 1954 to 1980 as being a consistently active period for tornadoes. After 1980, the amount of tornado activity drops off considerably. The 1980s and 1990s had relatively few tornadoes (with the conspicuous exception of 1992).

C. Frequency by Month

Figure 18 shows the distribution of tornado occurrences over the course of the calendar year. It displays the total number of tornadoes reported during each month since 1950. As expected, the greatest number of tornadoes occurs from mid-spring to mid-summer. The fact that May reports are much less than April and June is explained in the following section. There is a secondary peak, albeit quite small, in October representing strong autumn storm systems during the transition from the warm season to the cool season.

D. Frequency by Month, Without April 11, 1965 and April 3, 1974

In Figure 19, the tornadoes from April 11, 1965 and April 3, 1974 were removed. Despite the fact that only two days of data were removed from nearly 50 years' record, the exclusion of the Palm Sunday Outbreak and Super Outbreak made a dramatic difference. With the removal of the two outbreak events, there is a steadier rise to the peak activity in June, and the drop in activity between April and June is diminished. Tornado activity drops off rapidly after peaking in June.

E. Frequency by F-Scale

In Figure 20, we see the total number of each tornado strength reported. Strengths F0, F1, and F2 make up the distinct majority of the tornadoes in this area, with major tornadoes being exceedingly rare. Only one tornado of F5 strength has occurred, which was the Dunlap Tornado on the southeast side of Elkhart, Indiana on April 11, 1965. Of the 16 F4 tornadoes, 75% of them took place in either the Palm Sunday Outbreak (nine F4s) or the Super Outbreak (three). Therefore, it can be seen that violent tornadoes are quite rare in NWSO Northern Indiana's CWA.

F. Frequency by Time of Day

The pie chart (Figure 21) shows that nearly nine out of every ten tornadoes occur between noon and midnight. This implies that most of this area's tornadoes come from strong storm systems that occur in the heat of the day. Relatively few tornadoes are spawned by nocturnal convection, such as the mesoscale convective systems mentioned in Section 4.

G. Actual Number of Tornadoes per County

As seen in Figure 22, Berrien County has had the greatest number of tornado reports from within its boundaries, a total of 24. Nearby Marshall and Elkhart Counties had 23 and 22 tornadoes, respectively. All three counties are in the northwest part of the CWA. Marshall County shares a border with Elkhart County, and Elkhart County is only 12 miles away from Berrien County. Of the 24 counties that have had 11 or more tornadoes, 16 of them (67%) are in the northwest half of the CWA.

Many different factors contribute to the number of tornadoes reported for a given county. For example, Allen County (Indiana) and St. Joseph County (Indiana) rank fifth and seventh in total number of reported tornadoes, respectively. However, these counties are also the two most populous counties in the CWA. Hillsdale and Adams Counties are interesting in that they both have reported large numbers of tornadoes despite their small populations.

Another factor to consider is the size of the county. Allen County (Indiana) and Hillsdale County (Michigan) are the largest counties in the CWA and also have very high tornado totals. Blackford and Whitley Counties are quite small and have correspondingly low tornado totals. Anomalies of size versus tornado totals include Adams County that is relatively small yet has had a large number of tornadoes reported.

A third factor to consider is the availability of information from each county. Some counties are in frequent close contact with the NWS and provided many timely and accurate storm reports. Other counties have less-developed communications with the NWS, thus the NWS had fewer reports of tornadoes from those counties. Cass and St. Joseph Counties in Michigan and DeKalb County in Indiana are in close proximity to several counties with high tornado totals, yet these three counties had relatively low tornado totals. Perhaps the reporting networks in these counties were not fully developed or utilized. In other words, these counties may have had just as many tornadoes as their neighbors. Simply, these tornadoes were not reported to the NWS.

A final consideration is the effect on the statistics of the Palm Sunday outbreak and the Super Outbreak. Due to the large number of tornadoes that occurred on those two dates, the counties affected by the outbreaks showed higher numbers than counties left untouched on those two days. A comparison of the "Outbreak Counties" with the total number of tornadoes indicates that Marshall, Elkhart, Hillsdale, and Adams Counties were all affected by both outbreaks. The counties of Lagrange, Steuben, Grant, and Wells were also affected during both outbreaks and have near average tornado totals. This may indicate that these four counties normally do not have many tornadoes and their numbers have been inflated by the two outbreaks.

Berrien County is intriguing in that it has the highest tornado total, yet was not affected by either outbreak. Allen County (Indiana) also has a high tornado total but was not affected by the outbreaks. However, these counties are large in area and also have high populations, which likely contributed to the higher totals.

The slightly higher number of tornadoes in the northwestern section of the CWA may be related to mesoscale interactions with Lake Michigan, such as lake breezes. This would be particularly true in hard-hit Berrien County, Michigan and Saint Joseph, Elkhart, and Marshall Counties in Indiana.

H. Cold Air Funnels

Cold air funnels are often mistaken for tornadic funnel clouds but the two are quite different. Cold air funnels are funnel clouds that form in cold air masses generally from the late spring to the early autumn. They are not associated with true tornadic outbreaks and form under different conditions. Cold air funnels typically form when a cold air mass overlies a warmer air mass at the surface. The temperature profile is usually stable; however, very cold air aloft can decrease the stability. There is a higher frequency of cold air funnels near the coast of Lake Michigan (Cooley and Soderberg 1973). Moisture is often limited to the lower levels as it is advected off of the lake surface, and this may be why cold air funnels form in stratiform clouds or in thunderstorms with relatively low tops. Once the funnel forms, the more dense cold air aloft can drop below the base of the cloud and sometimes even reach the surface (however, this is rare). As stated earlier, cold air funnels form under different conditions than tornadoes do. Cold air funnels may produce gusty winds at the surface if they lower to near ground-level but wind speeds usually do not reach severe criteria (greater than 50 knots). Minor damage may occur; although, it is usually localized since cold air funnel last only a short time.


Lightning is not one of the criteria used for issuing severe thunderstorm warnings; however, lightning is the cause of more deaths and injuries annually than almost any other weather-related phenomenon. Only flash floods and river floods combined kill more people than lightning (Curran et al. 1997). According to the NOAA publication Storm Data from 1959 through 1997, 29 deaths and 116 injuries have been attributed to lightning in KIWX's CWA. This is an average of approximately three deaths every four years and just under three injuries per year. In addition, during the same 39-year period, another 117 lightning-related events have caused property damage (mainly due to fire) or some type of inconvenience to the public (such as electrical power outages). It's entirely possible that there were other occurrences of damage caused by lightning that were not reported. These 262 lightning reports translate into an average of almost seven events per year during the 39 years.

As can be expected, lightning damage is mainly a warm season effect. Almost 60% of the total of number of lightning events for the period (156) has occurred in June and July. Lightning damage reports generally start to be received in April, reach a peak in June and July, and usually end in September. Of the total of 262 events that were listed for the CWA in Storm Data from 1959 through 1997, only ten reports of lightning damage occurred during the six months from October through March (Figure 23). The reason for this is probably from a combination of things. More thunderstorms occur during the warmer half of the year, and people tend to spend more time outdoors.

Of the 262 lightning events recorded in Storm Data, 179 of them listed an actual time of occurrence (as opposed to "afternoon", "evening", "p.m.," etc.). During the afternoon between 12 noon EST and 6:00 p.m. EST, 44% (79 events) occurred. In addition, nearly 2/3 (112 events) occurred between 12 noon and 8:00 p.m. (Figure 24). As might be expected, lightning seems to be a phenomenon that occurs mainly during the afternoon and early-evening. No lightning-related damage was reported between 12 midnight and 1:00 a.m. and only 18 events (10%) were reported for the six hours between 12 midnight and 6:00 a.m. EST.

Deaths and injuries that are attributable to lightning generally are isolated occurrences. One remarkable event, however, occurred in Van Wert County, Ohio on September 5, 1977. Twenty-five people were injured while watching a parade at the Van Wert County Fair when lightning struck a fence they were leaning on. This is the largest number of people injured in the CWA due to a single lightning-related occurrence. During another lightning-related event on August 5, 1988, eight firefighters were injured while fighting a fire caused by a lightning strike in White Pigeon, Michigan.

Lightning may be seen with thunderstorms of any strength. Thunder is caused by the rapid collapse of the superheated air column associated with a lightning bolt. Also, just because lightning is not seen, this does not mean that a thunderstorm will not cause severe damage. Lightning may be obscured by the heavy rainfall associated with a particular thunderstorm but large hail and strong downburst winds can still cause extensive damage reaching severe criteria.


A. River Flooding

The headwaters and upper reaches of many rivers lie within the boundaries of NWSO Northern Indiana's Hydrologic Service Area (HSA). The HSA is generally contiguous with the CWA; however, it encompasses entire river basins. Perhaps the Wabash River is the largest and most notable. It begins in extreme west-central Ohio and flows across northern Indiana before turning southwest in Huntington County and turns south in Fountain County to flow to the Ohio River.

Two St. Joseph Rivers flow through the HSA and start within ten miles of each other in central Hillsdale County (Michigan). One flows on a winding path to the west, passing through extreme southern Michigan before entering Indiana in Elkhart County. It then flows southwest into St. Joseph County before turning abruptly North. (South Bend, Indiana gets its name from its location at this southernmost bend in the river.) It then flows to the northwest through Berrien County (Michigan) and empties into Lake Michigan . The other St. Joseph River flows southwest through Williams and Defiance Counties in northwest Ohio and then through southeast De Kalb and northern Allen Counties in Indiana.

The St. Mary's River forms in Auglaize County (Ohio) at Grand Lake St. Mary's. The river then flows northwest through Mercer and Van Wert Counties in Ohio. The St. Mary's completes its trek by flowing through Adams County and southern Allen County in Indiana.

The St. Joseph joins the St. Mary's at Ft. Wayne to form the Maumee River. The Maumee flows to the northeast and enters Lake Erie at Toledo. As can be imagined, the confluence of the St. Joseph and St. Mary's Rivers at Ft. Wayne can cause serious flooding problems, especially when both rivers are in flood upstream from Ft. Wayne. This area, however, has been inhabited for at least the last 200 years. The Miami Indians had a settlement here named Kekionga in the late-18th Century. American soldiers, under General Anthony Wayne, built a fort here in the early 19th Century (Fort Wayne). The Continental Divide is located to the southwest of this area and provided a portage for early Native American and European river travelers to move between the St. Mary's-Maumee River system to the northeast and the Wabash River basin to the southwest.

The Maumee River has a gradual slope, falling only around 200 feet from its beginning at Ft. Wayne to its mouth at Lake Erie, a distance of 100 miles. Two other large rivers join the Maumee in NWSO Northern Indiana's CWA. The Tiffin River flows through northwest Ohio and enters the Maumee from the north. A larger river, the Auglaize, drains a large area of west-central Ohio and enters the Maumee from the south. These two rivers meet the slow flowing Maumee at Defiance, Ohio. If these rivers are already high due to heavy rainfall or snowmelt flooding can become a problem in the city of Defiance and also upstream.

River flooding is a common occurrence throughout the CWA, especially in the spring when locally heavy rainfall can combine with snowmelt. Still-frozen ground can also impede runoff. River floods can affect huge areas and rivers can remain out of their banks sometimes for weeks. Floods generally affect a larger area of our CWA than any other weather-related phenomenon, with the possible exception of snowstorms. It is imperative that we become aware of how floods affect the region and what steps we can perform to predict the causes and behavior of floods. The adverse and tragic effects of flooding can possibly be lessened but rivers will probably always be impossible to totally control. We have to learn to live with the rivers in our midst.

Floods are known to have occurred across the region periodically throughout the 19th Century. However, river stage data has only been kept for approximately the last 100 years. The Organic Act of 1890 gave the NWS (then Weather Bureau) the responsibility for not only weather forecasting and storm warnings but also "the display of weather and flood signals for the benefit of agriculture" (Stallings 1991).

One of the first major floods to affect the current HSA after the Weather Bureau assumed flood warning responsibilities was also one of the major floods of the 20th Century. This flood occurred in March 1913 after heavy rain fell across much of the southern Great Lakes region including Indiana and Ohio. River crests during this flooding event remain the record levels for many locations, including the gauge locations in Table 1.

The next major flooding in NWSO Northern Indiana's HSA took place throughout the winter and spring of 1950. Many river-level records were set, especially in southern Lower Michigan. The record crest for the Little River (at Huntington, Indiana) occurred on January 4, 1950. The river crested at 20.00 feet, five feet above flood-stage. Three months later, the St. Joseph River in southwestern Michigan experienced flooding. On April 5, the St. Joseph (at Niles, Michigan) crested at 15.10 feet (4.1 feet above flood-stage). At Mottville, on April 27, the river crested at 10.76 feet (2.76 feet above flood stage). These crests remain the record stages for these three locations. The Maumee River (at Defiance, Ohio) crested 7.5 feet above flood at 17.50 feet. This is only the fourth highest crest the Maumee has experienced at this location. The record flood crest of 1913 is the all-time highest, and this 17.50 foot crest has since been surpassed by flood crests from 1982 and 1985.


River flood data for the record 1913 flood.


Flood Stage
Stage (Ft)


Maumee River
Ft. Wayne, IN FTWI3 15 26.10 3/26/1913
Defiance, OH DEFO1 10 26.00 3/26/1913
Napoleon, OH NAPO1 12 25.00 3/27/1913
St. Mary's River

Decatur, IN

DCRI3 15 26.50 3/26/1913
Blanchard River

Ottawa, OH

OTTO1 23 33.30 3/13/1913
Wabash River
Bluffton, IN BLFI3 10 21.00 3/25/1913
Wabash, IN WABI3 12 28.70 3/26/1913
Peru, IN PERI3 20 28.10 3/26/1913
Logansport, IN LGNI3 17 25.30 3/26/1913
Mississinewa River

Marion, IN

MZZI3 10 19.20 3/27/1913
Eel River

North Manchester, IN

NOMI3 8 15.50 3/27/1913


In the autumn of 1954, flooding occurred in the Kankakee River basin in northwestern Indiana. The Yellow River, a tributary of the Kankakee, was particularly hard hit. At Plymouth, Indiana, the Yellow River crested at 17.13 feet on October 12, 5.13 feet above flood stage. This remains the record crest for Plymouth.

Extensive flooding occurred across Indiana and Ohio during February 1959. At many locations this flood was exceeded only by the great 1913 flood. At Monticello, Indiana and Huntington, Indiana these 1959 flood levels have not been equaled. Table 2 lists some of the highest crests during this event.

Flooding occurred across the region during May and June 1978. Records were set in southern Michigan and northwest Ohio. The Tiffin River (at Stryker, Ohio) crested at 16.36 feet on May 23 (flood stage is 11.0 feet). This remained the record height for only four years, until the incredible flooding during the spring of 1982. Nottawa Creek in southern Michigan also set a record that was not broken until the flooding in spring 1989. On June 29, 1978, Nottawa Creek (at Athens, Michigan) crested at 6.47 feet (1.47 feet above flood stage).


River flood data for the record flood of 1959


Stage (Ft)
Stage (Ft)


Tippecanoe River
Monticello, IN MCX I3 9 16.70 2/10/1959
Wabash River
Huntington, IN HNTI3 20 23.20 2/10/1959
Wabash, IN WABI3 12 24.44 2/11/1959
Peru, IN PERI3 20 22.60 2/11/1959
Logansport, IN LGNI3 17 19.69 2/11/1959
Maumee River
Defiance, OH DEFO1 10 15.80 2/13/1959
Napoleon, OH NAPO1 12 19.50 2/11/1959
St. Joseph River
Montpellier, OH MONO1 12 14.20 1/31/1959
Blanchard River

Ottawa, OH

OTTO1 23 29.72 2/11/1959


A flood which rivaled the 1913 flood in destructiveness occurred in Marc, 1982. After a record snowfall that winter, heavy rain melted the deep snowpack. An underlying sheet of ice and frozen soil combined to keep the rapidly-forming runoff from entering the ground (DOC 1982). The runoff quickly filled the rivers of northeastern Indiana, northwestern Ohio, and southern Michigan. The city of Ft. Wayne was particularly devastated with damage estimated at $51 million (Glatfelter and Chin 1988).

This flood set the record crest at four different locations in northern Indiana and northwestern Ohio and was the second highest crest ever at eight other locations. Table 3 lists the record crests set during this event.

Three years later, an almost equally extensive flood occurred. However, damage was not as great due to precautions taken in many areas since the 1982 flood. The city of Ft. Wayne, for example, added levees to sections of downtown, and an ALERT rain gauge network was installed to help the NWS and other users monitor the levels of rainfall across the area. At many locations throughout the southern Great Lakes region river levels from this 1985 event were the second or third highest on record. At three river gauge locations (Winamac, IN; Montpellier, OH; and Nottawa, MI) flood levels were the highest ever. The stage that was recorded at Davis, Indiana, on the Kankakee River in early March, as the second highest level ever recorded at this location. Table 4 lists some of the crest data from this event.

A flood in June 1989 was confined mainly to the St. Joseph River basin in southern Michigan and northern Indiana. The rivers running through Athens, Michigan and Three Rivers, Michigan recorded their highest levels ever while the St. Joseph River (at Mottville, Michigan) rose to its second highest level. Table 5 lists some levels recorded during this event.


River flood data for the flood of 1982.


Stage (Ft)
Stage (Ft)


Maumee River
New Haven, IN NHVI3 15 25.49 3/17/1982
St. Joseph River

Newville, IN

NVLI3 11 17.96 3/17/1982
Tiffin River

Stryker, OH

STRO1 11 18.36 3/15/1982
Elkhart River
Goshen, IN GSHI3 7 11.94 3/14/1982
St. Joseph River
Three Rivers, MI TRVM4 7 10.69 3/21/1982
Elkhart, IN EKMI3 24 27.91 3/21/1982
Niles, MI NILM4 11 14.97 3/21/1982
Prairie River

Nottawa, MI

NOTM4 4* 6.12 3/20/1982
Yellow River

Plymouth, IN

PLYI3 12 16.37 3/16/1982
St. Mary's River

Decatur, IN

DCRI3 15 24.40 3/14/1982
Maumee River

Ft. Wayne, IN

FTWI3 15 25.93 3/17/1982

Defiance, OH

DEFO1 10 20.50 3/15/1982
* Bankfull stage; no flood stage established.



Flood data for the flood of 1985.


Stage (Ft)
Stage (Ft)


Tippecanoe River
Winamac, IN WINI3 10 15.40 2/20/1985
St. Joseph River
Montpellier, OH MONI3 12 17.40 2/25/1985
Prairie River

Nottawa, MI

NOTM4 4* 6.30 2/26/1985
* Bankfull stage; no flood stage established







Flood Stage (Ft) Record Stage


St. Joseph River
Three Rivers, MI TRVM4 7 10.52 2/27/1985
Mottville, MI MOTM4 8 9.70 2/28/1985
Elkhart, IN EKMI3 24 27.05 2/27/1985
Niles, MI NILM4 11 14.96 2/25/1985
Elkhart River

Goshen, IN

GSHI3 7 11.87 2/24/1985
Kankakee River

Davis, IN

DAVI3 10 13.52 3/05/1985
Tippecanoe River
Monticello, IN MCXI3 9 16.10 2/24/1985
Wabash River

Bluffton, IN

BLFI3 10 16.20 2/25/1985
Eel River

Adamsboro, IN

ADMI3 10 16.20 2/25/1985
Little Wabash River

Huntington, IN

LRHI3 15 19.50 2/25/1985
St. Mary's River

Decatur, IN

DCRI3 15 24.31 2/25/1985
Maumee River

Ft. Wayne, IN

FTWI3 15 24.55 2/27/1985

Defiance, OH

DEFO1 10 18.50 2/25/1985



Flood data for the June 1989 flood event.


Stage (Ft)
Stage (Ft)


St. Joseph River
Three Rivers, MI TRVM4 7 11.80 6/03/1989
Mottville, MI MOTM4 8 10.41 6/04/1989
Elkhart, IN EKMI3 24 26.16 6/04/1989
Niles, MI NILM4 11 13.30 6/04/1989
Nottawa Creek

Athens, MI

ATNM4 5 7.85 6/02/1989
Prairie River

Nottawa, MI

NOTM4 4* 5.7 4 6/03/1989
* Bankfull stage; no flood stage established


The flooding event at the end of December 1990 was more extensive. Flooding was reported throughout the HSA. The river-levels at some locations were the second and third highest ever recorded. Some of the highest readings at their respective locations are listed in Table 6.


Flood data for the December 1990 flood.


Flood Stage
Record Stage


Prairie River
Nottawa, MI NOTM4 4* 5.98 1/01/1991
Kankakee River

Davis, IN

DAVI3 10 13.48 12/30/1990
Yellow River

Plymouth, IN

PLYI3 12 15.30 1/01/1991
Tippecanoe River
Ora, IN ORAI3 10 15.00 12/31/1990
Elkhart River

Goshen, IN

GSHI3 7 11.03 12/30/1990
St. Joseph River

Montpellier, OH

MONO1 12 17.00 12/31/1990
Maumee River
Ft. Wayne, IN FTWI3 15 23.90 1/01/1991
New Haven, IN NHVI3 15 23.76 1/01/1991
Defiance, OH DEFO1 10 17.20 1/01/1991
Napoleon, OH NAPO1 12 17.13 1/01/1991
Tiffin River

Stryker, OH

STRO1 11 16.10 12/31/1990
Eel River

North Manchester, IN

NOMI3 8 14.81 12/30/1990

Adamsboro, IN

ADMI3 10 11.75 12/31/1990
Mississinewa River

Marion, IN

MZZI3 10 15.64 12/31/1990
* Bankfull stage; no flood stage established


B. Flash Flooding

River flooding can affect a large part of an entire river basin, but flash flooding can be just as destructive in the relatively smaller area it affects. A flash flood is defined as a "flood which is caused by heavy or excessive rainfall in a short period of time, generally less than six hours." (DOC 1990).

The topography across NWSO Northern Indiana's HSA is fairly level with no areas of great relief. This is due mainly to the effects of the last Ice Age (Wisconsinan Glaciation). At its greatest extent, reached almost to the present location of the Ohio River. Debris carried south by the glacier was deposited across the Great Lakes region. These sediments consist mainly of surface till. Depending on the source region of the debris is composed mainly of clays and loams. This soil, very beneficial for agriculture, can also affect the extent that rainfall can be absorbed. For example, in the Maumee River basin the soil in northeastern Indiana is mainly sandy in character and drains fairly well. Much of the rest of the basin is covered with clayey and loamy soils that drain more slowly (IDNR 1996). This can cause a flash flood problem during heavy rainfall events.

Flash floods across the region are not usually caused by a large change in elevation. During the warm part of the year flash floods usually occur as a result of intense rain showers falling on soil that is not able to absorb the precipitation fast enough to keep it from running off. Over urban areas, the large expanses of concrete will not allow any precipitation to reach the soil underneath. If the rainfall is too heavy for the storm drains to handle, the runoff then can create a danger to people living nearby. Driving becomes hazardous, especially in underpasses, and the fast-flowing waters can damage or destroy buildings and personal property.

During the winter and spring, flash floods can be caused from snow rapidly melting and frozen ground does not absorb the snowmelt. Ice-covered ground can aggravate this situation even more. Also during winter and spring, ice jams on rivers can cause water to back up and overflow the banks.

Flash flooding can be a serious problem in the HSA. Even though flash floods do not cover the large areas or last for a long period of time as river floods do, they can affect people just as much.


Although not meeting the criteria for the issuance of severe thunderstorm or tornado warnings, winter weather can adversely affect more people simultaneously across the CWA than almost any other type of weather (with the possible exception of widespread flooding). Winter-type precipitation has been experienced in the CWA during the months of September through May. By mid-December, during a typical winter, ice forms on the lakeshore of all the Great Lakes including Lake Michigan. With the exception of Lake Erie, the entire surface-area of the lakes does not become completely ice-covered except during the coldest winters. This shoreline ice serves to protect the coastline from wave erosion since the ice cover can extend into the lake hundreds of feet from shore.

A. Snow

The climate of the Great Lakes region is characterized by fairly cold, snowy winters. The lakes have a significant influence on the surrounding shores year-round (Sousounis 1995) but this can be especially seen during the winter. Average annual snowfall amounts vary across the CWA. In the lake effect snow regions in the lee of southeastern Lake Michigan, totals during the winter average 80 inches and have exceeded 150 inches in some active years. Farther inland, in northeastern Indiana and northwestern Ohio, average yearly snowfall totals are around 30 inches.

B. Synoptic-scale Systems

Snowfall generally occurs over the region as a result of one of two processes (or sometimes from a combination of the two). The first process is when a large low pressure system moves south of the region, and snowfall occurs to the north and west of the counter-clockwise circulation as cold air is advected south. This can happen at any time throughout the colder half of the year. Intense low pressure systems, as they move through the region, can drop more than a foot of snow across an area in a period of less than 24 hours. The strong winds often accompanying these systems can add the hazard of blowing snow. White-out conditions during these blizzards can make travel impossible. A synoptic event such as this can affect most of the northeastern U.S.

The Great Lakes are known for their powerful storms. Storms on the southern Great Lakes can affect the weather far inland in NWSO Northern Indiana's CWA. These systems can become especially potent during autumn when the lake surfaces are still generally ice-free. The relatively warm lakes can add moisture and instability to the air moving across them. This can be especially hazardous for shipping and for residents living along the shores. And, to complicate matters, it can be more difficult for a ship to maneuver around a large storm on the lakes than it would be on the open ocean (Eichenlaub 1979).

Some of the most memorable storms have occurred during autumn, before lake ice closes the shipping season. A powerful storm on Lake Huron on November 9 and 10, 1913, damaged or sunk dozens of ships. The Armistice Day Storm on November 11, 1940, occurred mainly on Lake Michigan and some mariners felt that this storm was as powerful as the 1913 storm. Other storms became famous because of the almost "unsinkable" ships that were lost with all or most of their crews. On November 18, 1958, the great bulk freighter Carl D. Bradley sunk. Huge ore freighters have also been lost such as the Daniel J. Morrell on November 29, 1966 and the Edmund Fitzgerald on November 10, 1975. The effects of these storms were felt, not only on the lakes themselves but for hundreds of miles inland.

Some significant snow events:


Jan. 1918

The snowiest January on record in Ft. Wayne until 1982. Heavy snowfalls at the beginning of the month and again two weeks later. A total of 25.4 inches of snow fell throughout the month. The station monthly summary listed 13.5 inches of snow still on the ground at the end of the month.
Nov. 1950

A Thanksgiving snowstorm enabled this to become the snowiest November on record in Ft. Wayne. The U.S. Weather Bureau monthly record stated that the "snow storm on November 25 and 26 blocked roads in every direction." A total of 8.0 inches of snow fell during both days, with 2.5 inches more occurring over the following two days. At South Bend a total of 10.1 inches fell over the four-day period.
Mar. 1964

A late-season storm dumped over a foot of snow across portions of the southern Great Lakes region from March 9 - 12. A total of 16.4 inches of snow was reported at Ft. Wayne over the four-day period (12.6 inches fell just on March 10). This remains the second snowiest March on record at Ft. Wayne (behind March 1912). However, South Bend only received a total of 5.2 inches during this event.
Dec. 1973

Heavy snow fell across the southern Great Lakes region December 19and 20. During the storm 14.0 inches of snow fell at Ft. Wayne, helping to make it the second snowiest December on record (behind December 1914). South Bend received 12.9 inches during the event.
Nov. 1977

At South Bend, over a period of four days (25 - 28), a total of 24.4 inches of snow fell, making this the snowiest November on record (14.6 inches fell on November 25 alone). The snowfall season of 1977-78 was to become the snowiest on record at South Bend with a seasonal total of 172.0 inches.
Jan. 1978

A blizzard during the last week of January paralyzed most of the Great Lakes region. Snow fell from January 25 - 27. Snowfall totals of up to 40 inches fell in parts of northern Indiana. Drifts of up to 25 feet were reported in northern Ohio. Cars were abandoned on highways when motorists were stranded. Storm Data listed this blizzard as the worst of the century in Ohio. This was the snowiest January on record in South Bend with 86.1 inches of snowfall recorded. Ft. Wayne recorded its third snowiest January ever with a total of 25.3 inches.
Jan. 1982

A blizzard hit northern Illinois and Indiana January 9 - 10 and a second snowstorm hit northern Indiana and Michigan January 15 and 16. The month ended up being the snowiest January on record in Ft. Wayne with 29.5 inches of snow recorded.
Jan. 1987

January became the fifth snowiest January on record at Ft. Wayne when 17.6 inches fell during the month. A snowstorm on January 9 and 10 dropped over half a foot of snow over northern Indiana. During a second storm from January 17 - 20 up to a foot of snow fell across portions of northern Indiana and northwest Ohio. At Ft. Wayne 9.1 inches fell during the event and at South Bend 9.9 inches were recorded. During a third snowstorm January 22 - 24, 11.8 inches fell at South Bend.


C. Lake Effect Snow

The second of the two processes mentioned above is lake effect snow. This is a mesoscale process which involves a cold air mass moving across a relatively warmer water surface. This cold air picks up water vapor from the surface and deposits it as snow on the lee (downwind) shore of the body of water. With longer fetches (the distance the air mass travels across the lake) the air can take up more water vapor and thus generate larger amounts of lake effect snow. Lake effect snow bands are generally not more than 20 miles wide and range from 50 to 100 miles in length (Peace and Sykes 1966).

Wind direction plays a large role in what areas of the CWA will see lake effect snow. A northerly wind trajectory straight down the long axis of Lake Michigan will affect portions of northeast Illinois, northwest Indiana, and extreme southwest Michigan. In KIWX's CWA the counties of Berrien in southwestern Michigan and La Porte and St. Joseph in northwest Indiana will be particularly affected. A change in wind direction to the northeast can cause the lake effect bands to miss the CWA entirely and impact northeast Illinois and southeast Wisconsin. If the wind backs too northwesterly, the counties mentioned above are still affected but the snow bands will reach farther inland to the east. A change to westerly or west-southwesterly winds may allow the snow bands to miss Indiana entirely and blanket much of southwest Lower Michigan. This is why it can be such a challenge to predict the areas to receive snowfall and the amount.

The probability of a "pure" lake effect snow event occurring (in the absence of a synoptic system adding to the snowfall) increases as certain criteria from the following list are met:

With the addition of new technology and methodologies forecasters are better able to predict what locations will be most affected by a lake effect snow event, and even which counties in our CWA that are at most risk for experiencing heavy snowfall. Due to the mesoscale nature of the phenomenon, predicting LES trajectories beyond 6 hours can be fraught with difficulty (Waldstreicher et al. 1996). Areas that are affected by LES events are dependent on many factors (moisture, instability, etc.). Change in wind direction or vertical shear can cause a LES band to intensify, dissipate, or transition from single to multiple bands (or vice versa) over an area in less than an hour. A lake effect snow event has to be continually monitored to see how it may be evolving.

Orography can affect the amount of snow that ultimately falls. Numerical simulations involving changes in elevation around Lake Michigan revealed that snowfall rates could be enhanced by several hundred percent with large elevation increases near the lake shore (Hjelmfelt 1992). This can be a factor in our counties in northwest Indiana and southwest Michigan. The area around South Bend, for example, is almost 150 feet higher in elevation than the surface of Lake Michigan. This rise, over a distance of just 25 miles, can add to the upward motion of the already unstable air passing over the lake. The problem is not as significant in our northwest Ohio counties since they range from 30 to 60 miles from Lake Erie to the northeast. In addition, the topography across northwest Ohio slopes upward more gradually so the effect of a change in elevation would not be as pronounced.

The location of the KIWX WSR-88D radar in Kosciusko County (Indiana) positions it close enough to both Lake Michigan and Lake Erie to detect the relatively low-topped lake effect or lake enhanced bands that affect the region. As was found during the LES study conducted by the NWS during the mid-1990s, the use of satellite imagery can extend the effective range of the WSR-88D even farther (Carter et al. 1996).

The Great Lakes (with the exception of Lake Erie) do not usually freeze over completely during the winter. With this large open water surface, evaporation can continue throughout most of the winter. Significant lake effect snowfalls can occur within 50 miles of the shore. This can be seen when comparing the average snowfall totals from Fort Wayne and South Bend. Fort Wayne is approximately 100 miles from both Lake Michigan and Lake Erie. South Bend, by contrast, is located only around 30 miles southeast of Lake Michigan. Because of its nearness to the lake South Bend averages 81.8 inches of snowfall each winter while Fort Wayne, farther inland, averages 34.8 inches (from 1960-1990 30-year normals). Table 7 lists the average monthly snowfall and record monthly snowfall for Fort Wayne and South Bend.


Average monthly snowfall & record monthly snowfall - September through May

Fort Wayne

South Bend
Average Snowfall

Record Snowfall

Average Snowfall
Record Snowfall
September 0.0 0.0 0.0 1.2
October 0.4 8.0 1.0 8.8
November 3.2 14.1 8.6 30.3
December 7.9 20.3 21.0 41.9
January 8.3 29.5 22.8 86.1
February 8.3 16.9 16.0 35.1
March 5.3 19.5 9.7 33.9
April 1.4 11.7 2.7 14.0
May Trace Trace 0.0 0.6


Snow spotters add immeasurably to the ability of forecasters to predict the location of lake effect snow bands. A snow spotter network is being developed in NWSO Northern Indiana's CWA. According to the Weather Service Operations Manual, "NWS offices are encouraged to recruit and maintain winter weather spotter networks" (DOC 1992). The addition of snow spotter reports will greatly augment the available radar and surface observation data.

D. Freezing Precipitation

Snow is not the only type of winter precipitation that can paralyze the Great Lakes region. Freezing precipitation, rain and drizzle, can deposit an inch or two of ice on exposed surfaces. Roads, sidewalks, power lines, and vehicles can become covered with ice during a freezing precipitation event. Freezing precipitation results when snow falls through a layer of air that is above the freezing point of water and then into a second layer that is below freezing. Snowflakes melt into raindrops as they fall into this warmer layer, and these drops then fall into another atmospheric layer that is below freezing. If the raindrops freeze before hitting the ground they are termed ice pellets (commonly referred to as sleet). If the raindrops freeze only after coming into contact with the earth's surface (or with objects at the surface such as automobiles, fences, etc.) they are defined as freezing rain.

NWSO Northern Indiana's CWA can expect to experience a freezing precipitation event once per winter, on average. Some memorable events since 1970:


March 24-26, 1978
Ice-storm across northern Illinois, Indiana, southern Michigan and northern Ohio.
Winter of 1989
Ice storms occurred across southern Michigan in each of three months (Jan. 5-6, Feb. 2, and March 3-4).
February 14-16, 1990
Freezing rain with 2 to 4 inches of accumulation in northwest Indiana and an inch in northwest Ohio.
March 12-13, 1991
Ice-storm combined with 6 to 12 inches of snow in northern Indiana and northern Ohio.


E. Extreme Cold Temperatures

The southern Great Lakes region has a continental climate that is modified somewhat by its proximity to the lakes. However, the lakes do not exert enough influence to keep very cold outbreaks from occurring, especially when the lakes are mostly frozen over. During a typical winter, overnight low temperatures will fall to below zero Fahrenheit (0°F) at least once and many winters will see low temperatures at a particular location fall below zero on three or four different nights. Rarely, does the temperature not rise above zero during the day except during the most severe arctic outbreaks. Average and record low temperatures for two sites, Fort Wayne and South Bend are given in Table 8.

The typical cold outbreak scenario involves strong high pressure over central Canada. Northerly winds east of the high advect cold air into the Great Lakes region. If there is snow cover already across the region this will help to cool the arctic air mass even more.

Extremely cold temperatures can adversely affect people in many ways. Some effects are a nuisance, such as an automobile that fails to start. Other effects can be life-threatening, as when people are stranded in stalled cars in sub-freezing weather. Table 8 lists the average temperatures and the record low temperatures in degrees Fahrenheit (°F) for Fort Wayne and South Bend.


Average monthly temperatures & record monthly low temperatures - September through May

Fort Wayne

South Bend
Average Record Low Average Record Low
September 64.9 29 63.9 29
October 52.9 19 52.9 20
November 41.3 -1 40.9 -7
December 28.6 -18 28.9 -16
January 22.9 -24 23.3 -22
February 25.9 -18 26.4 -17
March 37.6 -10 37.4 -13
April 49.2 7 48.7 11
May 60.2 27 59.4 24


Some notable cold outbreaks (temperatures in °F):


Feb.  1899

Incredible cold -outbreak in mid-February (11-14) that set low temperature records throughout the eastern U.S. (at many locations in the deep south these temperatures remain the lowest recorded since records have been kept). Unofficially, the temperature dropped to -20°F at South Bend.
Jan. 1918

An extremely cold month with below zero temperatures recorded on eight days during the month at Ft. Wayne (including a period of five days from January 17 - 21 when the temperature dropped below zero each day). Temperature at Ft. Wayne fell to -24°F January 12.
Jan. 1943
All-time record low temperature set for South Bend at -22°F. Low at Fort Wayne reached -13°F January 20.
Feb. 1951
Temperature dropped to -17°F at both South Bend and Ft. Wayne on February 2. This remains the record low for February at South Bend.
Jan. 1972

Temperature fell to below zero on three consecutive nights (14, 15, and 16) at Ft. Wayne. Low reached -18°F January 15 and then dropped to -19°F on January 16. At South Bend the low reached -10°F on January15 and -19°F on January 16.
Jan. 1977
Extreme cold conditions lasted for most of the month across the Great Lakes region.
Feb. 1982

Second week of February was extremely cold across the region. Temperature on the February 10 fell to -18°F at Ft. Wayne and to -13°F at South Bend. A foot of snow cover contributed to the radiational cooling.
Dec. 1983

Bitterly-cold temperatures were experienced during the last two weeks of December. Readings dropped to below zero on nine nights at Ft. Wayne and on eight nights at South Bend. The coldest night was December 24 when -15°F was recorded at South Bend and -16°F at Ft. Wayne.
Jan. 1985

Below zero temperatures were recorded at both Ft. Wayne and South Bend on three consecutive nights (19, 20, and 21). Ft. Wayne's temperature January 20 fell to -22°F. South Bend recorded a low of -21°F on this date, which just missed tying the all-time record low of -22°F.
Dec. 1989

Below normal temperatures recorded through most of the month. On December 22 the temperature at South Bend dropped to -13°F. Ft. Wayne's all-time lowest December reading was recorded on December 22 when the temperature dropped to -18°F.



During the period of record for this study (1950-1997), adverse weather was reported in every month of the year. Significant severe convective events were most numerous in June and July. Winter season events included a myriad of activity including synoptic-scale systems which brought blizzard conditions and ice storms. Lake-effect snows were responsible for a dramatic increase in snowfall in the northwestern portion of the CWA. Although reporting was subject to area and population values, some conclusions can still be drawn with regard to specific occurrences of significant weather phenomena.


The authors wish to thank Paul Janish and John Hart of the Storm Prediction Center (SPC) for their assistance in procuring a copy of SVRPLOT. Appreciation is also extended to Gary Beeler and David Eversole of NWSO Mobile, Alabama for their helpful advice concerning the installation of SVRPLOT on Northern Indiana's personal computers. We would particularly like to thank Mike Sabones, Meteorologist in Charge, and Jane Hollingsworth, Warning Coordination Meteorologist, at Northern Indiana for offering many welcome suggestions on how to improve this document.


Carter, G.M., T.L. Salem, Jr., and M.A. Wool, 1996: The Use of New Technology for Enhanced Snow Forecasts in the Vicinity of Lakes Erie and Ontario. Preprints, 15th Conference on Weather Analysis and Forecasting, Norfolk, AMS (Boston) 569p.

Cooley, J.R., and M. E. Soderberg, 1973: Cold Air Funnel Clouds. NOAA Technical Memorandum NWS CR-52. DOC, NOAA, NWS Central Region, Scientific Services Division,Kansas City, MO, 29pp.

Curran, E.B., R.L. Holle, and R.E. Lopez, 1997: Lightning Fatalities, Injuries and Damage Reports in the United States, 1959-1994. NOAA Technical Memorandum NWS SR-193. DOC, NOAA, NWS Southern Region, Scientific Services Division, Fort Worth, TX, 64pp

DOC, NOAA, NWS, 1982: Midwest Floods March 13-20, 1982. Natural Disaster Survey Report, 4pp.

____________, 1990: Flood/Flash Flood Watch and Warning Program. Weather Service Operations Manual, Chapter E-20, 4pp.

____________, 1992: Winter Weather Warnings. Weather Service Operations Manual, Chapter C-42, 31pp.

____________, 1995: Severe Local Storm Watches, Warnings, and Statements. Weather Service Operations Manual, Chapter C-40, 8pp.

Doswell, C. A., III, and D.W. Burgess, 1993: Tornadoes and Tornadic Storms: A Review of Conceptual Models. The Tornado: Its Structure, Dynamics, Prediction, and Hazards (Geophysical Monograph 79), Amer. Geophys. Union, 161-172.

Eichenlaub, V., 1979: Weather and Climate of the Great Lakes Region. Univ. of Notre Dame Press, 282-304.

Glatfelter, D.R., and E.H. Chin, 1988: Floods of March 1982 in Indiana, Ohio, Michigan, and Illinois. U.S. Geological Professional Paper 1467. United States Government Printing Office, 1pp.

Grazulis, T.P., 1993: Significant Tornadoes 1680-1991. Environmental Films, St. Johnsbury, VT, 283pp.

Hart, J.A., 1993: SVRPLOT: A New Method of Accessing and Manipulating the NSSFC Severe Weather Data Base. Preprints, 17th Conference on Severe Local Storms, St. Louis, AMS (Boston), 40-41.

Hjelmfelt, M.R., 1992: Orographic Effects in Simulated Lake-Effect Snowstorms over Lake Michigan. Mon. Wea.Rev., 120, 376-377.

Hough, J.L., 1958: Geology of the Great Lakes. University of Illinois Press, 35-44.

Indiana Dept. of Natural Resources, 1996: Water Resource Availability in the Maumee River Basin, Indiana. IDNR, Division of Water, 32-57.

Johns, R.H., 1984: A Synoptic Climatology of Northwest-Flow Severe Weather Outbreaks. Part II: Meteorological Parameters and Synoptic Patterns. Mon. Wea. Rev., 112, 449-464.

____________, and W.D. Hirt, 1987: Derechos: Widespread Convectively Induced Windstorms. Wea. Forecasting, 2, 32-49.

____________, and C. A. Doswell III, 1992: Severe Local Storms Forecasting. Wea. Forecasting, 7, 588-612.

NOAA, NCDC 1959-97: Storm Data. National Climatic Data Center, Asheville, NC 1-39, No's. 1-12.

Peace, Jr., R.L., and R.B. Sykes, Jr., 1966: Mesoscale Study of a Lake Effect Snow Storm. Mon. Wea. Rev., 94, 495pp.

Sousounis, P.J., 1995: Multi-Scale Weather in the Great Lakes. First Symposium on Michigan Weather and Forecasting. SE Mich. Chapter of the AMS, 1pp.

Stallings, E., 1991: Flood Forecasting and Drought Prediction by the National Weather Service. National Water Summary 1988-89 - Hydrologic Events and Floods and Droughts, U.S. Geological Survey, 117-122.

Wakimoto, R.M. and J.W. Wilson, 1989: Non-supercell Tornadoes. Mon. Wea. Rev., 117, 1113-1140.

Waldstreicher, J.S., T.A. Niziol, and E.A. Mahoney, 1996: Lake Effect Snow Forecasting - 1985 to 1995: A Decade of Scientific and Public Service Advancements. Preprints, 15th Conference on Weather Analysis and Forecasting, Norfolk, AMS (Boston), 559pp. is the U.S. government's official web portal to all federal, state and local government web resources and services.