STREAMFLOW CONDITIONS, FLOODING, AND LOW FLOWS


From The Changing Illinois Environment: Critical Trends, Volume 2: Water Resources , Technical Report of the Critical Trends Assessment Project

INTRODUCTION

Streams in Illinois serve a variety of different uses. They serve as water supply sources for municipal, industrial, and rural needs, and receive wastewater from water treatment plants, thereby incorporating it into the environment. They also provide for recreation, and serve as habitat for fish, water birds, and numerous other types of aquatic life. Often these uses conflict with each other, particularly on streams with high use and/or during drought conditions. As the use of Illinois streams intensifies, it is an increasing challenge to maintain all of the stream functions. Alteration of the flow in streams by natural or artificial means can either facilitate or hinder these uses. Streamflow trends may also signal potential impacts to water quality and changes in the stream ecosystem.

Five factors were identified as having possible, long- term impacts on the amount of flow in streams:

Climate variability Changes in land use (urbanization, reforestation, and removal of wetlands) Reservoir construction and operation Water use (stream withdrawals and discharges) Channelization and other changes in drainage

The objective of this investigation is the analysis of changes in Illinois streamflow conditions, with particular emphasis in relating how the above factors affect the flow regimes of the streams. The identification of trends in streamflow conditions is of great value in analyzing other environmental trends because of the potential impact of flow quantity on water use conflicts, water quality, aquatic habitat, sedimentation and erosion, and stream morphology. In the material presented below, no attempt has been made to further review these potential impacts.

METHODOLOGY

Streamgage Data

All streamgaging records in the state with continuous flow records extending 50 years or more were examined to identify long-term trends in streamflow conditions. Certain locations in the state do not have streamgages with 50-year records, most notably the metropolitan area of northeastern Illinois. For these locations, additional stations with flow records of approximately 40 years or more were added to the dataset. This resulted in a dataset of 79 gauging stations (table 1). Out of this list, individual subsets were created for analyzing the effect of each of the five factors affecting streamflow conditions. For example, in analyzing the impacts of climatic change, it is necessary to choose streamgages with flow records that are relatively untouched by other factors. The streamgages used in analyzing each factor are presented later in this report.

Flow Parameters Evaluated

Each streamflow record contains a minimum of 14,000 values of average daily streamflow. It is impractical to use these data directly to estimate long-term trends in streamflow conditions, primarily because the daily values fluctuate tremendously and are dominated by seasonal differences. Instead, various summary statistics (termed herein as "flow parameters") are computed for each year of the record, and trend analysis is per- formed on the annual series of each flow parameter. The following seven flow parameters were used in the analysis. All flow values are given in cubic feet per second (cfs):

Annual average flow 7-day low flow 7-day high flow Average winter flow (December-February) Average spring flow (March-May) Average summer flow (June-August) Average fall flow (September-November)

Statistical Analyses

Three types of statistical analyses were employed to identify trends in the annual series for each flow parameters: 1) Kendall Tau-b trend analysis, 2) linear regression analysis, and 3) autocorrelation and spectral density analyses. All three analyses used a null hypothesis test to determine the significance of trends at a 95 percent level of confidence. The results of all three analytical procedures were fairly similar, thus only the Kendall analysis results are reported here.

Table 1 presents the correlation coefficients computed using the Kendall analysis for all 79 gauging stations. The Kendall Tau-b analysis (Kendall, 1975) produces a rank correlation coefficient, which indicates the strength of the trend relative to the overall variability of the flow parameter being analyzed. A coefficient of zero indicates absolutely no trend. Negative values indicate a decreasing trend, and positive values indicate an increasing trend. The strength of the trend increases as the absolute value of the coefficient approaches 1.0 (the maximum correlation). If the natural variability of the flow parameter is particularly great, as with high flows and floods, it may be more difficult to identify a significant trend.

IMPACTS OF CLIMATE VARIABILITY

Weather and climate are the primary driving forces that determine the amount and distribution of streamflow. Variability is a natural aspect of the climatic and hydrologic processes. Annual variability in climate, particularly in precipitation, may result in periods of drought or extended high streamflow. Figure 1 illustrates the considerable variation that can occur with both annual precipitation amounts and the resulting streamflow. The 11-year moving averages of both the precipitation and streamflow are also shown to illustrate their long-term variation. The moving averages indicate that above- and below-normal conditions can persist over a period lasting as long as two decades. When using short streamflow records, it can be particularly difficult to differentiate between this natural variance and a long-term change in overall climatic conditions. For this reason, it is desirable to use long-term streamflow records to examine trends that may be associated with climatic change.

Previous Studies

Ramamurthy et al. (1989) and Singh and Ramamurthy (1990) examined the increases in average annual flows and peak flows observed on the Illinois River for the period 1941-1985. These studies found that average annual flow on the Illinois River had increased by 20 to 25 percent since 1970, and that annual peak flows had increased about 50 percent. The number of days having high-flow conditions doubled during that time. These studies concluded that the higher flows were caused by concurrent increases in precipitation amounts throughout much of the Illinois River basin.

Changnon (1983) examined trends in the number of flood events and the duration of flood flows for 11 large watersheds in Illinois over the period 1921-1980. Changnon concluded that the number of flood events and flood days generally increased over that period, particularly in the northern and eastern portions of Illinois. Changnon also found an increase in the number of heavy rainfall events during the summer (May to August), which is believed to be one of the causes for the increase in summer floods. Gauging Records Used in the Analysis

The streamgage records used to analyze the impact of climate variability on hydrology should come from watersheds relatively unaffected by human-induced impacts, including withdrawals, return flows, reservoirs, detention storage, and major land-use changes. Additional gage selection criteria, developed by Slack and Landwehr (1992) for evaluating impacts of climate variability, are the length of the gagging record, its general level of accuracy, and the relative lack of missing or estimated data. Slack and Landwehr identified 36 streamgage records from Illinois as appropriate for the study of climate impacts on streamflow, using a mini-mum record length of 20 years. For this analysis, a more stringent requirement of 40 years of record was applied, thus eliminating 9 of those 36 records. The gauging record for the Rock River at Afton, Wisconsin, was included since this stream enters Illinois slightly downstream of the gage. Another gauging record, the Fox River at Algonquin, is not without minor influences on its flow, but was included because it provides an important long-term record of average flow conditions in northeastern Illinois. The resulting 29 stations used in the analysis are identified in table 2.

The Kendall correlation coefficients for each streamflow parameter and each of the 29 stations are included in table 1. An examination of these coefficients indicates that most values have a positive correlation, thus we can conclude in general terms that flow conditions are increasing throughout the state. However, most of these coefficients are less than the threshold needed to identify individual trends at the 95 percent level of confidence. Only in certain portions of the state are the trends significantly large to pass this threshold. Figure 2 presents the locations of those stations where statistically significant trends in flow conditions were identified. A description of the flow changes at selected stations and by region follows. Changes in Mean Streamflow

Most stations throughout Illinois have experienced above-average flow conditions in the last 25 years. This does not necessarily indicate a trend in streamflow conditions, however. Hypothesis tests having a 95 percent confidence level are used to statistically determine which increases are significantly different from zero, thereby suggesting an actual trend rather than natural streamflow variability.

Eleven gauging stations in northern Illinois experienced significant increases in mean streamflow over their period of record, as indicated by the hypothesis tests. The hatched area in figure 2a identifies the general region of impact. Records for the remaining 18 locations throughout the state either have no change in mean flow or an amount of change that is not sufficiently different than zero to conclude that a trend exists.

Magnitude of Changes. Figure 3 presents the annual mean discharge for seven streamgages in Illinois with record lengths greater than 75 years. The moving average displays the variation of above- and below-normal conditions throughout the record. All of the watersheds experienced below-normal flow in the 1930s and 1950s, and above-normal conditions in the 1970s and early 1980s. However, two streams--the Fox River at Algonquin and the Kankakee River at Momence--have experienced a particularly high mean flow rate since the late 1960s.

Table 3 compares the percent difference in the mean flow observed at all 29 stations in the last 26 years (1966 to 1991) to the common 51-year period from 1941 to 1991. Also shown is the percent difference from the longer 77-year record of 1915 to 1991.

Kankakee River vicinity. The Kankakee River, in particular, appears to display steady increases in mean flow over its period of record. All three long-term gauging records in the Kankakee River basin show significant increases in mean flow over the last 75 years. The flow in the last 25 years, when compared to that at the start of the period of record, is more than 40 percent greater--an average increase of 0.5 percent per year. The relationship of this increase to changes in the average precipitation of the Kankakee River watershed is examined later.

Northern Illinois. All watersheds in northern Illinois experienced a 13 to 20 percent increase in mean flow in 1966-1991 above the long-term average, with the exception of the northwestern corner of the state--on the Pecatonica and Apple Rivers. The Kishwaukee River at Belvidere and Perryville and the Fox River at Algonquin had particularly high flows during this period, 20 percent above average.

Central Illinois. The watersheds in central Illinois have experienced a 10 to 15 percent increase in mean flows since 1966. While this is a considerable increase, it is not sufficient for the analysis to detect a definite trend in mean flow conditions.

Southern Illinois. Southern Illinois watersheds have experienced little change in mean flow conditions over the last 75 years. Impact of Period of Record on Analyzing Trends. Examination of the long-term records in figure 3 indicates that the identification of a trend can be partly dependent on the period of record being examined. For example, an evaluation of the Monticello, Seville, and Freeport gage records between 1935 and 1985 would suggest that these locations have experienced steady increases in mean flow. When the full period of record at these stations is examined, however, the existence of a trend is less apparent. Most of the gauging stations being analyzed have a period of record from about 1940 to 1991, or starting from a period where below-normal flows were observed and including the period of above-normal conditions from 1972-1985. It may be expected that some of these shorter records will display an increasing trend when no long-term trend is actually present.

Table 4 lists the Kendall trend coefficients developed for 12 long-term gauging stations in Illinois. Trend coefficients for the entire period of record are compared with partial records containing a shorter period of 50 continuous years. Trend coefficients that exceed the 95 percent level of confidence are presented in bold italics. The coefficients with the highest absolute values represent the strongest trends, whether positive (increasing trend) or negative (decreasing trend). Examination of these values indicates that the shorter 50-year records are much more likely to have greater coefficients than the entire period of record. The exceptions to this are the stations on the Kankakee River, where the trend coefficients for the overall record are as high or higher than any of the partial records. This supports the contention that a significant long-term trend exists on this river.

Table 4 also shows that several gages displayed negative coefficients, indicating decreasing trends in flow, prior to 1973. It is quite possible that 25 years ago there may have been as much concern over decreasing flow conditions as there is today over increasing trends. This brings up an additional question of whether cyclical trends exist on these streams. The possible existence of cyclic trends is briefly examined in the upcoming section: What is the Nature of the Trends? But the length of the gauging records under examination is generally not sufficient to identify long-term cyclical trends.

For many of the stations listed in table 4, the trend correlation for the 50-year records ending in 1982 are significantly higher than for either the entire period of record or any of the other 50-year partial records. Some of the previous analyses concerning trends in streamflow (Changnon, 1983; Ramamurthy et al., 1989) also used streamflow records that extended to the early and mid-1980s. These earlier studies may have results that more strongly suggest trends than the results presented herein, simply because of the period of record used in the analysis. In a similar manner, it should be expected that analyses conducted in the future, using longer gauging records, will also have different results but should present a clearer understanding of long-term trends.

Changes in Low Flows

All 11 stations having an increase in average flow also have an increasing trend in low flows (see figure 2b). Of the remaining 18 stations outside the region, only one has a significant increase in low flows: the Apple River near Hanover.

Figure 4 displays the annual series of 7-day low flows observed at five of the long-term gauging stations. The Fox River at Algonquin is not included because its low flows are noticeably affected by an upstream reservoir. The dark line in each of the plots shown is the 11-year moving median of the low flows. Low flows in the central and southern parts of Illinois (Embarras River, Spoon River, and Sangamon River) were above normal during the 1970s but otherwise show little change over time.

The low flows on the Kankakee River show only a slight trend since 1924, but were consistently below normal from 1916 to 1923. This period of below-normal low flows on the Kankakee River coincides with the period when channelization of that river was being completed (Knapp, 1992), and therefore may have some human- induced impacts. Figure 4 illustrates that the Pecatonica River has experienced an increase in low flows since 1980. But unless future decades show further increase, this is not considered a trend and, in fact, the Kendall hypothesis tests do not identify a significant trend.

Changes in the Number of Low Flow Occurrences. The analysis used to identify trends provides only a comparison of the general magnitude of the annual series of 7-day low flows, and it does not distinguish those years with truly dry conditions from those years where low flows were not severe (or where a low-flow condition did not really exist). In analyzing extreme conditions, such as low flows, it is also necessary to look at the frequency at which these extreme events occur.

Table 5 lists the average number of "significant" low flow events that occurred throughout the gauging records. The "significant" low flows shown in this table are those with magnitudes so low that they occur on average only once in five years. On average, each decade should be expected to have two of these events. This table indicates the change in their frequency of occurrence.

Kankakee River vicinity. In and near the Kankakee River watershed, the frequency of low flows is remarkably consistent from the mid-1920s until 1970. But in the 20- year period 1971-1990, only one severe low-flow event occurred (the 1988 drought). This suggests a definite change in low-flow frequency in the last two decades.

Northern Illinois. The northern Illinois gauging records suggest a recent tendency toward less frequent low flows. The occurrence of low flows in this part of Illinois tends to aggregate in two decades, the 1930s and the 1950s, suggesting a significant interannual dependence between the low flows. The decade of the 1920s, at the beginning of the gauging period, and the period 1971-1990 had significantly low occurrences of severe low flows. This fluctuation in low flow frequency suggests that if a long-term trend toward less frequent low flows exists, it is more likely to be cyclical rather than linear in nature. But it is not possible to objectively conclude that a long-term cyclical trend exists given the period of record.

Central and southern Illinois. Low flows in the central and southern portions of the state have occurred with the greatest frequency during three decades: the 1930s, 1950s, and 1960s. There is no apparent trend that the frequency of severe low flow events is changing. In fact, low-flow frequency during the most recent decade, the 1980s, is very indicative of the long-term norm.

Changes in High Flows and Flood Peaks

Only seven stations are identified as having significant increases in high flows, and six of these stations are located in or near the Kankakee River basin in northeastern Illinois (figure 2c). These increased high flows on the Kankakee River may also be the primary cause of flooding increases observed farther downstream on the Illinois River, reported by Ramamurthy et al. (1989) and Singh and Ramamurthy (1990).

Figure 5 displays the annual series of 7-day high flows observed at seven long-term gauging stations. The Kankakee, Fox, Spoon, and Sangamon River gages all experienced increases in high flows, though trends were identified for only the first two of these gages. The stations in southern Illinois (Embarras River and Cache River) show no increase in high-flow conditions. The 11-year moving average of high flows for the Pecatonica River at Freeport shows a consistent decrease in high flows throughout its period of record. This Kendall trend coefficient for the high-flow series at Freeport is -0.149, which approximates, but does not exceed the threshold used to identify significant trends. (The Kendall analysis was performed on the annual high-flow series, which has a high degree of variability. For these types of series, the trend must be particularly strong to have the coefficient exceed the 95 percent level of confidence.) In a study of a tributary of the Pecatonica River, Potter (1991) suggests that decreasing floods may be the result of land-use changes, particularly soil and water conservation.

Changes in the Number of Flood Events and Flood Days. Table 6 provides the number of significant flood events at several gauging stations by decade. For the purpose of this analysis a significant flood event is one where the expected frequency of the event is less than approximately once every two years. Table 7 provides the number of flood days by decade at the same stations, i.e., the number of days where the average daily flow exceeds the significant flood level. These values are patterned after similar ones presented in Changnon (1983) and were extended to include the years 1981-1990. Changnon (1983) identified an increase in flood events in eastern Illinois and an increase in flood days throughout much of the northern and eastern portions of the state. With ten additional years of data (1981-1990), the existence of a consistent trend in the number of flood events is less clear, and the observed increases were not sufficient for the statistical tests to demonstrate any trends. The number of flood days appears to have increased for the Kankakee River near Wilmington and the Sangamon River at Monticello, but decreases are noted for both the Pecatonica River at Freeport and the Cache River at Forman.

Changes in Seasonal Flows

The annual series of the mean flow rate for four sea- sons (March-May, June-August, September-November, and December-February) were examined to identify changes in streamflows by season. Trends in seasonal flows were identified at 11 gauging stations (identified in figure 2d by nonzero values). Ten of these stations are located in the same region identified as having increases in both low and average flows. All 11 stations experienced increased flow during the fall (September-November).

Increases in mean flow rate over all four seasons were displayed by two gauging records: the Kankakee River at Momence and at Wilmington. The two gages on the Kishwaukee River, at Belvidere and Perryville, experienced increased flows for all seasons but the summer (June-August), while the Rock River had increased flows during the winter, and the Fox River at Algonquin had increased flows during the summer.

Comparison of Streamflow Increases and Precipitation Increases

Changes in the Illinois climate were examined in the Atmospheric Resources volume. The period from 1966-1983 experienced greater precipitation and streamflow throughout Illinois than did any other period this century. Precipitation and streamflow records indicate, however, that other wet periods were experienced earlier this century and in the nineteenth century, and standard statistical analysis does not indicate that the latest trends are significantly different. In the past, these wet trends have not lasted more than about 20 years. Were it not for the possible impacts of increased concentrations in greenhouse gases, it might be expected that in future years the climate and streamflow conditions would return to more normal conditions. In most locations throughout the state, the average streamflow in the period since 1985 has, in fact, been much closer to the long-term average.

Ramamurthy et al. (1989) and Singh and Ramamurthy (1990) related the increases in average precipitation and average streamflow in the Illinois River watershed. These studies indicate that average precipitation over northern Illinois has increased by 10 to 14 percent for the period since 1965. These precipitation increases resulted in a 20 to 25 percent increase in the mean flow. This finding is similar to the results of Knapp and Durgunoglu (1993), who simulated the impacts of increased precipitation on a central Illinois watershed using a rainfall-runoff computer model, and found that a 10 percent increase in precipitation produces a 25 percent increase in average runoff.

Northern Illinois. Figure 6 compares the streamflow measured on the Kishwaukee River at Belvidere and the rainfall observed at Marengo. The precipitation increase coincides almost exactly with increased streamflow. The increases in the 1965-1980 precipitation and streamflow over the long-term average are 12 and 24 percent, respectively.

Kankakee River Vicinity. Figure 7 compares the precipitation measured at Kankakee with the streamflow for the Kankakee River near Wilmington. The comparison suggests that a 10 percent increase in average precipitation roughly correlates to a 25 percent increase in average streamflow.

Central and Southern Illinois. Locations in central and southern Illinois also experienced above-normal precipitation during the years 1966-1983 (for example, see figure 1). But the increase is less than 6 percent, and the average streamflow for this period was not substantially greater than at all other times during the period of record.

What is the Nature of the Trends?

The identification of a trend using a statistical test does not necessarily indicate the nature of that trend. For example, an increasing linear trend suggests that changes have been consistent over the period of analysis, and possibly that additional increases will continue. Three hypotheses are briefly examined:

The trends are linear in nature and suggest similar increases/decreases in the future. The trends are cyclic in nature, periodically rising and falling over the period of record. When a change in streamflow conditions occurs, it is essentially instantaneous, resulting in a distinctly different set of average flow conditions (step trend).

Three long-term gages that have shown trends in streamflow conditions were examined: the Kankakee River at Momence, the Fox River at Algonquin, and the Rock River at Afton, Wisconsin. For each gauging record linear, cyclic, and step-change functions were fitted to the annual average flow. Parameters for all trend functions were identified by least squares regression techniques. These trend functions are illustrated for each of the three stations in figure 8. The correlation coefficients between the observed annual flows were estimated to determine which trend function most closely approximates the observed changes. The computed correlations are presented in table 8.

Table 8 indicates that in all cases the step trend functions more closely match the series of annual mean discharges. This suggests a distinct change in northern Illinois flows since 1970, although they are not continuing to increase. In fact, the linear trend has the worst correlation for all stations except the Kankakee River at Momence. Figure 8 also shows that since 1985, the mean streamflow at the Afton and Algonquin gages has been closer to the long-term mean than to the 1965- 1985 conditions. It is unclear whether this indicates a return to "more normal" conditions. The mean flow at the Kankakee River at Momence has remained above average.

Discussion and Conclusions

Most of the northern third of Illinois shows significant increases in both average streamflow and low flows. Much of northern and central Illinois also received significant increases in average precipitation in the period 1966-1983. The locations that experienced increased average and low streamflows generally experienced precipitation increases >10 percent; therefore the streamflow increases are likely caused by changes in precipitation. Less significant increases in streamflow are also noted in central Illinois, which experienced less change in precipitation. Average precipitation and streamflow since 1984 is close to the long-term average. The locations that show an increase in average flow uniformly show increases in low flows and flows during the fall season. High flows have not generally increased, however. The Kankakee River watershed in northeastern Illinois continues to experience anomalous behavior, having continuous increases in precipitation and all streamflow parameters.

The statistical tests identified increasing trends for high flows and flood peaks for the Kankakee River basin. Increases in the number of flooding days have been observed during the last two decades for the Kankakee River. But for most of Illinois, statistically significant increases in high flows and flood peaks do not appear. This is somewhat contrary to the findings of Changnon (1983), who used ten years' fewer data.

IMPACTS OF LAND-USE CHANGE

Urbanization

In the process of urbanization, numerous alterations to a watershed may occur that will cause change in streamflow conditions. The addition of impermeable surfaces, including paved roads, parking lots, roof surfaces, etc., reduces the amount of infiltration and decreases evapotranspiration, thereby increasing both average runoff and storm runoff. More recently, stormwater detention has been added in many watersheds to reduce flood peaks. In most cases, the decreases in infiltration (and hence groundwater recharge) result in a lower groundwater table and less streamflow during dry periods. But for certain urban watersheds, the cumulative effect of a variety of return flows-- caused by lawn watering, sump pumping, cooling systems, or other small discharges--may increase low flows.

In the early 1950s numerous gages were located in the Chicago metropolitan area to measure flow from urbanizing watersheds. Table 9 lists the stations in water-sheds that have undergone the greatest amount of urbanization since the gages were installed. All water- sheds have experienced changes associated with urbanization that affect the flow. Of particular note are the construction of stormwater detention reservoirs, which alter high flows and flooding, and return flows from wastewater treatment facilities, which greatly increase low and medium flows in streams. The presence of these major modifications in each of these watersheds is noted in table 9. The impacts of stormwater detention and wastewater effluents are individually analyzed later in this section.

Analysis of the impacts of urbanization is further complicated in that all of these watersheds are located in northeastern Illinois, which has experienced increased streamflow conditions resulting from climate variability.

Average Flows. No change in the average flow rate has been detected beyond that which may be caused by the impacts of climate change.

Low Flows. The annual series of low flows were examined for the four gauging records listed in table 9 that are not impacted by return flows. Increases in low flows are common, but not universal among these urban streams. This may be partially explained by return flows from lawn watering. The urban area also produces large amounts of wastewater, which must eventually be discharged into the streams. The effect of these return flows is examined in the section on Impacts of Water Use.

Flooding and High Flows. Drainage improvements, including storm sewers and channelization, are designed to expedite the flow of water away from urban land. In recent decades, stormwater detention has been added to urban areas to reduce the impact of downstream flooding resulting from these changes. The effects of all these changes on streamflow are well documented in many urban hydrology texts. Of the gages listed in table 9, only three are not affected by stormwater detention (Flag Creek, Tinley Creek, and Long Run). None of these three gage records show an increasing trend in flood peaks, although the Tinley Creek gage displays an increasing trend in the 7-day high flow.

Stormwater Detention. Most urban watersheds in northeastern Illinois have stormwater detention facilities to reduce flood peaks in the watershed. Table 9 identifies six urban watersheds in which stormwater detention reservoirs have been constructed. Trend analysis of the flood peak series indicates that five of these six watersheds have an increasing trend in peak flows. Figure 9 shows that the Addison Creek gage has experienced consistent increases in flood peaks, despite the construction of several detention reservoirs in the watershed. Only Midlothian Creek at Oak Forest displays a reversal in the increasing flood peaks (figure 10).

Reforestation/Deforestation

A considerable amount of forested area in Illinois was converted to farmland more than a century ago, prior to the installation of any streamgages. The trend over recent decades is a small increase in the amount of woodland in Illinois, particularly in the northern and southern extremes of the state. The amount of forest in a watershed must change significantly for these changes to be observed in the streamflow. A 20 percent change in overall runoff may be necessary to separate a trend induced by reforestation from the natural variability in the streamflow record. Since most of the gauging stations in Illinois are located on large streams, it is not possible to evaluate the impact on streamflow of small-scale reforestation.

Experimental catchment studies on the hydrologic impact of deforestation have been conducted in many parts of the nation (e.g., Douglass and Swank, 1972; Patric, 1973; Anderson et al., 1976; TVA, 1955, 1973). All of these studies indicate that deforestation results in increases in average streamflow, low flows, and flooding levels. The amounts of increase vary considerably and depend upon the type of forest cover and various watershed characteristics.

The increase in average flow results primarily from the reduction in watershed evapotranspiration. Increases in flow are particularly apparent the first year after deforestation, before other vegetative growth has a chance to become established. The decreased evapotranspiration also results in higher soil moisture values in unforested areas, which further results in greater baseflow (groundwater flow to streams) during dry periods (Patric, 1973). Flood volumes and peak flow rates often increase after deforestation. However, both Douglass and Swank (1972) and Anderson et al. (1976) indicate that this is usually a result of soil compaction and reduced infiltration associated with the land clearing and development that follows forest harvesting, and not by the removal of trees. Fewer experimental data are available on the impacts of forest regrowth on streamflow, but the few existing studies indicate that reforestation has a reverse impact and causes decreases in the average flow, low flow, and flooding (Anderson et al., 1976). It may take considerable time to reestablish the same flow conditions that existed prior to the loss of the forest, however.

Removal of Wetland Areas

Illinois has lost an estimated 85 percent of its original wetlands (Dahl, 1990). The impact of the removal of these wetlands on streamflow is not clear. It is often accepted that wetlands reduce streamflow peaks and increase low flows. A review of the scientific literature by Demissie and Khan (1993), however, indicates that this generalized concept may not always be true. Previous studies on the impacts of wetlands are often conflicting: some show that the presence of wetlands causes decreases in peak flows (Novitzki, 1982; Ogawa and Male, 1983), while others suggest an increase or no change in peak flows (Moklyak et al., 1972; Skaggs and Broadhead, 1982). For Illinois, Demissie and Khan (1993) indicate peak flow and floodflow volume are lower in watersheds that have a high percentage of wetland area; but they caution that this relationship could be influenced by other factors. Demissie and Khan also reported increases in low flows with decreasing wetland area, which is contrary to the generally accepted viewpoint. The contribution of other factors to these increases in low flows were not examined.

IMPACTS OF RESERVOIR CONSTRUCTION AND OPERATION

Major reservoirs can produce considerable changes in the flow characteristics of the streams on which they are located. Flood storage reservoirs in large rivers, in particular, are likely to greatly impact the streamflow regime (Kitson, 1984). But the extent to which flow changes occur is dependent upon the purpose, design, and operation of the reservoir. Peak flows and high-flow conditions will be diminished--the extent of the reduction depends on the storage-outflow characteristics of the reservoir. The frequency of medium-flow levels will also usually be slightly increased by a reservoir. Low flows downstream of Illinois reservoirs are usually decreased, although they can be increased if reservoir operation releases a minimum protected flow. The average flow from the reservoir will usually be only slightly reduced, primarily through evaporation from the surface of the reservoir. Knapp (1988) estimates that the loss in average annual flow is approximately 0.3 cfs for every square mile of surface area. Net seepage losses from the reservoir to ground water are usually considered to be small.

Reservoirs on Small Watersheds

Most reservoirs on small watersheds in Illinois have uncontrolled outflow, meaning that outflow occurs over the spillway and no gates are used to regulate the flow. There are no streamgage records for locations directly downstream of such reservoirs to provide examples of their impacts. Knapp (1988) used computer modeling of a reservoir water budget in developing a methodology to roughly estimate the new flow regime downstream. Reservoirs located on small watersheds often have a comparatively large capacity relative to the amount of water flowing into the reservoir, and in these cases there is a greater impact on local streamflow. Figure 11 provides an example of the simulated outflow from a reservoir on a small watershed. If a reservoir with uncontrolled outflow is also used for water supply, the low flows from it are reduced to an even greater extent. The impacts of the reservoir on high and medium flows are usually attenuated downstream of the reservoir--with sufficient distance, the impacts of the reservoir may be difficult to detect.

Major Reservoirs

Table 10 lists the eight streamgages in Illinois located either directly downstream of or sufficiently close to a major reservoir to be affected. Of the gages listed, the Decatur, Shelbyville, Carlyle, and Plumfield gages are located directly downstream of the reservoir.

Mean Flow. None of the gauging stations used in the analysis demonstrate a significant change in mean annual flows.

Low Flows. All of the eight gages listed in table 10 are located downstream of reservoirs that release a mini-mum protected flow with the exception of the Sangamon River at Decatur. Lake Decatur presents a unique example of possible impacts on low flows because the lake serves a large water supply function. Withdrawals from this lake for water supply far surpass the normal losses from reservoir storage that occur as a result of evaporation and seepage. The impact of Lake Decatur withdrawals on low flows is illustrated in figure 12, which provides the flow frequency relationship at the Decatur gage and the estimated inflow into the reservoir.

Increased low flows in the Fox River at Algonquin, shown in figure 13, result from changes in the operation policy, rather than from any inherent dam characteristics. The gage records from all the other reservoirs demonstrate an increase in low flows after dam construction, all as a result of protected flow releases from the reservoir. Trends in low flows on these streams are therefore subject to changes in policy, as compared to a physical cause-and-effect relationship. The low-flow increases on the Salt Creek near Rowell and the Kaskaskia River gages is moderate, <10 cfs. The gages on the Big Muddy River show substantial increases in low flows, shown in figure 14.

High Flows. Lakes Shelbyville and Carlyle, located on the Kaskaskia River, and Rend Lake on the Big Muddy River were all built primarily for flood control. The three gages directly downstream of these reservoirs (at Shelbyville, Carlyle, and Plumfield) all show significant reductions in flood flows, as illustrated for Lake Carlyle in figure 15. Two gages, at Vandalia and Murphysboro, are located approximately 50 miles downstream of Lake Shelbyville and Rend Lake, respectively. Both of the gages show only slight reductions in flooding, below a level that is detected by the Kendall analysis. In a similar manner, the Rowell gage, located 13 miles downstream of Clinton Lake, shows only a small reduction in flooding. Lake Decatur and the Fox Chain of Lakes, upstream of the Decatur and Algonquin gages, respectively, are not designed to provide much reduction in flooding and therefore show no trend.

Seasonal Flow. Reservoirs do not ordinarily produce a seasonal change in streamflows. However, Lake Shelbyville and Rend Lake change pool levels between winter and summer. The Shelbyville and Plumfield gages both show an increase in average winter flow, resulting from the storage release when the pool level at each of these lakes is lowered in early winter.

Impacts of Water Use (Stream Withdrawals and Discharges)

Streams serve both as sources of water supply and as receptors of wastewater from municipalities and industry (table 11). If the stream is used as a water supply source, the effluent discharge will occur a short distance downstream from the withdrawal. In this case, since the withdrawal and discharge are usually of similar magnitude, there will be an insignificant impact on the flow in the stream, except perhaps over the short distance between the two facilities. More commonly, the municipal water supply is obtained from ground water, and the wastewater discharge to the stream increases the flow. During normal and high-flow conditions, the magnitude of these discharges is small compared to the ambient streamflow. However, in most cases the effluent discharge is large compared to the ambient low-flow conditions in the river, and a significant increase in low flow occurs. In a few cases, there is an interbasin transfer of water, such that a water supply withdrawal reduces the flow in one stream, and returns the wastewater to a different stream.

Mean Flow

Mean flows are affected only with a sizable effluent into a relatively small watershed. The gauging stations that show such an increase are all located in northeastern metropolitan Illinois. Direct withdrawals from streams are typically much smaller than the mean flow, primarily because they are designed to withdrawal no more water than what is available during dry periods. As a result, the impact of withdrawals on mean flow were not detected. Low Flow

The Kendall analysis did not detect a statistically significant trend in low flows for the two stations downstream of withdrawals, the Vermilion River at Pontiac and the South Fork Sangamon River near Rochester. But in recent years, zero flows have occurred at these gages when they previously did not (see figure 16), suggesting a change in extreme low flows.

Twenty-six gauging stations downstream of return flows demonstrate trends in low flows. These include most stations in northeastern metropolitan Illinois, along with the Vermilion River near Danville, South Branch Kishwaukee River near Fairdale, Bureau Creek at Princeton, Sugar Creek near Hartsburg, Sangamon River at Riverton, and Henderson Creek near Oquawka. The location of all streams, gaged and ungaged, that likely have increased low flows caused by return flows is shown in figure 17. Other locations, such as Richland Creek near Hecker, are impacted by return flows but have not experienced an increase water use over the gage's period of record--thus they show no trend in low- flow conditions.

IMPACTS OF STREAM CHANNELIZATION

Lopinot (1972) reports that approximately 27 percent of the stream mileage in Illinois is channelized. The distribution of channelized streams in Illinois is summarized in Mattingly and Herricks (1991). Channelization is particularly common in the urban northeastern portion of Illinois, and locations where the natural drainage is relatively poor, particularly the east-central portion of the state. Impacts of stream channelization can include increased downstream flooding, a lowered water table, destroyed stream habitat, reduction in stream vegetation, and increased sedimentation downstream. No available research attempts to estimate the change in streamflow conditions in channelized areas, other than applied studies on localized flooding effects. Changes in the flow regime caused by channelization along an established stream are believed to be subtle. To detect the impacts of stream channelization on streamflow, it would be necessary to strategically place gages upstream and downstream of a channelized area. No such set of gages exists with which to conduct an analysis.

SUMMARY

Table 1 provides the results of the Kendall trend coefficients for all stations that were examined. The most common impact on streamflow appears to be that associated with climate change in northern Illinois. Climate change in this portion of the state has produced an increase in average and low flows in streams. There does not appear to be a significant, widespread increase in flooding from climate change, although increased flooding is observed throughout the Kankakee River basin. Water use appears to have the second largest impact on flows in Illinois, primarily from effluent discharges to streams. This results in an increase in low flows for numerous streams in the state. Many urban streams in northeastern Illinois have considerable increases in high flows, regardless of whether stormwater detention facilities are present.

REFERENCES

Alley, W.M. 1988. Using Exogenous Variables in Testing for Monotonic Trends in Hydrologic Time Series. Water Resources Research 24(11): 1955-1961.

Anderson, H.W., M.D. Hoover, and K.G. Reinhart. 1976. Forests and Water: Effects of Forest Management on Floods, Sedimentation, and Water Supply. U.S. Department of Agriculture General Technical Report PSW 18/1976, Pacific Southwest Forest and Range Experimental Station, Berkeley, CA.

Changnon, S.A. 1983. Trends in Floods and Related Climate Conditions in Illinois. Climate Change 5(4):341- 363.

Dahl, T.E. 1990. Wetland Losses in the United States, 1780s to 1980s. U.S. Department of the Interior, Fish and Wildlife Service, Washington, DC.

Demissie, M., and A. Khan. 1993. Influence of Wetlands on Streamflow in Illinois. Illinois State Water Survey Contract Report 561, Champaign, IL.

Douglass, J.E., and W.T. Swank. 1972. Streamflow Modification through Management of Eastern Forests. U.S. Department of Agriculture, Forest Service Research Paper SE-94, Southeastern Forest Experimental Station, Asheville, NC.

Kendall, M.G. 1975. Rank Correlation Methods. Charles Griffin, London, 4th edition..

Kitson, T. (ed). 1984. Regulated River Basins: A Review of Hydrological Aspects for Operational Management. UNESCO International Hydrological Programme, Technical Documents in Hydrology.

Knapp, H.V. 1988. Fox River Basin Streamflow Assessment Model: Hydrologic Analysis. Illinois State Water Survey Contract Report 454, Champaign, IL.

Knapp, H.V. 1992. Kankakee River Basin Streamflow Assessment Model: Hydrologic Analysis. Illinois State Water Survey Contract Report 541, Champaign, IL.

Knapp, H.V., and A. Durgunoglu. 1993. Evaluating Impacts of Climate Change on Midwestern Watersheds. Paper presented at the Hydroclimatology Conference of the American Meteorological Society 1993 Annual Meeting, Anaheim, CA.

Lopinot, A.C. 1972. Channelized Streams and Ditches of Illinois. IDOC Division of Fisheries. Special Fisheries Report #35.

Mattingly, R.L., and E.E. Herricks. 1991. Channelization of Streams and Rivers in Illinois: Procedural Review and Selected Case Studies. Illinois Department of Energy and Natural Resources, Report ILENR/RE-WR-91-01.

Moklyak, V.I., G.B. Kubyshkin, and G.N. Karkutsiev. 1972. The Effects of Drainage Works on Streamflow. Hydrology of Marsh-Ridden Areas, UNESCO Press, Paris, France, pp. 439-446.

Novitzki, R.P. 1982. Hydrology of Wisconsin's Wetlands. Information Circular 90, U.S. Geological Survey, Madison, WI.

Ogawa, H., and J.W. Male. 1983. The Flood Mitigation Potential of Inland Wetlands. Water Resources Research Center Report 138, University of Massachusetts, Amherst.

Patric, J.H. 1973. Deforestation Effects on Soil Moisture, Streamflow. and Water Balance in the Central Appalachians. U.S. Department of Agriculture, Forest Service Research Paper NE-259, Northeastern Forest Experimental Station, Upper Darby, PA.

Potter, K.W. 1991. Hydrological Impacts of Changing Land Management Practices in a Moderate-Sized Agricultural Catchment. Water Resources Research 27(5):845-855.

Ramamurthy, G.S., K.P. Singh, and M.L. Terstriep. 1989. Increased Duration of High Flows along the Illinois and Mississippi Rivers: Trends and Agricultural Impacts. Illinois State Water Survey Contract Report 478, Champaign, IL.

Singh, K.P., and G.S. Ramamurthy. 1990. Climate Change and Resulting Hydrologic Response: Illinois River Basin. In Watershed Planning and Action, Symposium Proceedings of the Conference on Watershed Management, ASCE Irrigation and Drainage Division.

Skaggs, R.W., and R.G. Broadhead. 1982. Drainage Strategies and Peak Floodflows. Paper No. 82-2054 presented at the summer meeting of the American Society of Agricultural Engineers at the University of Wisconsin, Madison, June 27-30.

Slack, J.R., and J.M. Landwehr. 1992. Hydro-Climatic Data Network (HCDN): A U.S. Geological Survey Streamflow Data Set for the United States for the Study of Climate Variations, 1874-1988. U.S. Geological Survey Open-File Report 92-129, Reston, VA. Tennessee Valley Authority (TVA). 1955. Influences of Reforestation and Erosion Control on the Hydrology of the Pine Tree Branch Watershed, 1941 to 1950. Technical Monograph 86, TVA Division of Water Control, Hydraulic Branch Unit.

Tennessee Valley Authority (TVA). 1973. Summary Report on the Upper Bear Creek Experimental Project. TVA Division of Water Control, Hydraulic Branch Unit. Table 1. Correlation Coefficients of the Kendall Trend Analysis

Flow parameter REGION Years of Mean 7-day 7-day Fall Winter Spring Summer Station ID Location record flow low flow high flow mean mean mean mean

SOUTHEASTERN ILLINOIS 3339000 Vermilion River @ Danville 63 0.1531 0.3601 0.1869 0.1562 0.1336 0.1060 0.0824 3343000 Wabash River @ Vincennes 61 0.1424 0.3033 -0.0396 0.1299 0.0881 0.0746 0.1842 3345500 Embarras River @ Ste. Marie 77 0.0376 0.0786 0.0595 0.0321 0.0642 0.0540 0.0226 3346000 North Fk Embarras @ Oblong 51 0.0824 0.0726 0.1357 0.1388 0.1451 0.0823 0.0180 3379500 Little Wabash River @ Clay City 77 0.0451 0.0449 0.0581 0.0034 0.0656 0.0533 0.0027 3380500 Skillet Fork @ Wayne City 63 0.0753 0.0460 0.0937 0.0353 0.0896 0.0364 0.0783 3612000 Cache River @ Foreman 67 0.0402 0.1030 0.0041 -0.0023 0.0692 0.0882 - 0.0113

NORTHWESTERN ILLINOIS 5419000 Apple River @ Hanover 57 0.0739 0.1857 0.0990 0.0990 -0.0388 0.0376 0.0326 5430500 Rock River @ Afton, WI 77 0.0849 - 0.0032 -0.0056 0.0638 0.1179 0.0232 0.0351 5435500 Pecatonica River @ Freeport 77 0.0191 0.1158 -0.1490 0.0649 -0.0581 0.0226 0.0492 5437500 Rock River @ Rockton 52 0.2398 0.2674 0.0196 0.2564 0.2172 0.1418 0.0830 5438500 Kishwaukee River @ Belvidere 52 0.3228 0.2737 0.1508 0.2790 0.2081 0.1629 0.2112 5439500 S Branch Kishwaukee River @ Fairdale 52 0.2715 0.3255 0.1855 0.2700 0.1931 0.1523 0.2278 5440000 Kishwaukee River @ Perryville 52 0.2836 0.2376 0.1222 0.2730 0.1936 0.1599 0.2368 5443500 Rock River at Como 61 -0.1945 -0.1322 -0.2741 -0.1650 -0.2579 -0.1126 -0.0874 5444000 Elkhorn Creek @ Penrose 52 0.2308 0.3176 0.0422 0.2745 0.0332 0.1644 0.1372 5446500 Rock River @ Joslin 52 0.2474 0.3004 0.0603 0.2564 0.2021 0.1176 0.1508 5447500 Green River @ Geneseo 55 0.1892 0.2020 0.1354 0.1407 0.0707 0.1111 0.0748 5466000 Edwards River @ Orion 51 0.1090 0.0955 0.0337 0.1200 0.0541 0.0871 - 0.0384 5469000 Henderson Creek @ Oquawka 57 0.0564 0.3286 0.0075 0.1404 -0.0163 0.0200 - 0.0614

KANKAKEE RIVER REGION 5520500 Kankakee River @ Momence 76 0.3137 0.2252 0.2688 0.2070 0.1846 0.2400 0.2442 5525000 Iroquois River @ Iroquois 47 0.2581 0.2174 0.2026 0.2488 0.1452 0.1674 0.0028 5525500 Sugar Creek @ Milford 43 0.1384 0.0639 0.1340 0.1916 0.1096 0.1118 - 0.0631 5526000 Iroquois River @ Chebanse 68 0.1703 0.2483 0.1651 0.1220 0.0711 0.1422 0.1414 5527500 Kankakee River near Wilmington 76 0.3558 0.2685 0.2618 0.2190 0.2175 0.2596 0.2856 5542000 Mazon River @ Coal City 52 0.1922 0.4188 0.1843 0.3035 0.1514 0.0588 0.0337

NORTHEASTERN ILLINOIS 5529000 DesPlaines River @ DesPlaines 51 0.3474 0.6424 0.0855 0.3710 0.2282 0.1561 0.2910 5531500 Salt Creek @ Western Springs 46 0.4956 0.8040 0.1382 0.5285 0.2580 0.1594 0.4377 5532500 DesPlaines River @ Riverside 48 0.4681 0.7262 0.0727 0.5089 0.3280 0.1755 0.3865 5536000 North Branch Chicago River @ Niles 41 0.3976 0.7462 0.1512 0.4122 0.2585 0.1561 0.3220 5536275 Thorn Creek @ Thornton 43 0.3134 0.5424 0.0853 0.3931 0.1140 0.1185 0.1562 5536290 Little Calumet River @ South Holland 44 0.2770 0.4684 0.1353 0.3721 0.0761 0.0846 0.1860 5539000 Hickory Creek @ Joliet 47 0.1656 0.2309 0.0657 0.2081 0.1489 0.0583 0.0305 5540500 DuPage River @ Shorewood 51 0.4055 0.6147 0.1059 0.3929 0.2345 0.1843 0.3349 5546500 Fox River @ Wilmot, WI 49 0.3588 0.4486 0.1412 0.3214 0.2670 0.1905 0.1752 5550000 Fox River @ Algonquin 76 0.1979 0.2843 0.0758 0.1467 0.1790 0.0702 0.2035 5552500 Fox River @ Dayton 67 0.3152 0.4079 0.2800 0.2393 0.1976 0.2130 0.2953

SMALL URBAN STREAMS 3337000 Boneyard Creek @ Urbana 42 -0.1716 - 0.3240 -0.0742 0.0233 -0.1274 -0.0720 - 0.1938 5528500 Buffalo Creek @ Wheeling 39 0.3171 0.4282 0.1606 0.4143 0.2389 0.1687 0.2173 5529500 McDonald Creek @ Mt. Prospect 39 0.2119 0.5647 0.1471 0.3738 0.2038 0.0499 0.1795 5530000 Weller Creek @ Des Plaines 41 0.0415 0.1487 0.2463 0.1878 -0.0293 -0.0805 0.0073 5532000 Addison Creek @ Bellwood 40 0.5462 0.4791 0.2103 0.4718 0.3795 0.3308 0.3641 5533000 Flag Creek @ Willow Springs 40 0.5487 0.7058 0.2051 0.4128 0.4769 0.3667 0.4128 5534500 North Br. Chicago River @ Deefield 39 0.3306 0.6273 0.2578 0.4008 0.2713 0.1822 0.2011 5535000 Skokie River @ Lake Forest 40 0.1410 -0.0445 0.2231 0.2359 0.1897 0.0923 0.0436 5535500 West Fk. North Br. Chicago River 39 0.4359 0.5932 0.2173 0.5007 0.3036 0.2173 0.3414 5536215 Thorn Creek @ Glenwood 42 0.4448 0.3146 0.1429 0.3752 0.2242 0.2149 0.2172 5536235 Deer Creek @ Chicago Heights 43 0.2824 0.1336 0.2735 0.3444 0.1783 0.1539 0.0166 5536255 Butterfield Creek @ Flossmoor 43 0.1694 -0.1266 0.1960 0.3333 0.0498 0.0498 0.0011 5536265 Lansing Ditch @ Lansing 43 0.1429 0.2985 -0.1849 0.3200 0.0410 0.0210 0.1783 5536340 Midlothian Creek @ Oak Forest 41 0.2561 0.3641 0.1268 0.3415 0.1171 0.1073 0.1342 5536500 Tinley Creek @ Palos Park 40 0.3974 0.5978 0.2359 0.3795 0.2410 0.2103 0.1846 5537500 Long Run @ Lemont 40 0.2538 0.6815 0.0487 0.2202 0.2256 0.1359 0.0513 5550500 Poplar Creek @ Elgin 40 0.3026 0.2578 0.3154 0.4051 0.2205 0.1410 0.1077 Table 1. Concluded

Flow parameter REGION Years of Mean 7-day 7-day Fall Winter Spring Summer Station ID Location record flow low flow high flow mean mean mean mean

ILLINOIS RIVER 5543500 Illinois River @ Marseilles 52 0.2525 -0.0211 0.1582 0.1814 0.1219 0.1393 0.0987 5568500 Illinois River @ Kingston Mines 52 0.2685 0.3145 0.2383 0.3213 0.2051 0.1554 0.0996 5586100 Illinois River @ Valley City 51 0.1576 0.2490 0.1420 0.1843 0.1372 0.0682 - 0.0274

CENTRAL ILLINOIS 5554500 Vermilion River @ Pontiac 49 0.1616 0.1472 0.1565 0.2364 0.1429 0.0391 - 0.0867 5555300 Vermilion River @ Leonore 60 0.1642 0.1398 0.1818 0.1268 0.0859 0.0871 0.0894 5556500 Big Bureau Creek @ Princeton 55 0.1623 0.3222 -0.0263 0.2040 0.0330 0.1030 0.0182 5567500 Mackinaw River @ Congerville 47 0.1656 0.1130 0.1711 0.1415 0.0786 0.0657 - 0.1748 5569500 Spoon River @ London Mills 49 0.1360 0.1738 -0.0119 0.2177 0.0255 0.0697 - 0.0867 5570000 Spoon River @ Seville 77 0.1237 0.1193 0.0793 0.0390 0.0232 0.1415 0.0232 5572000 Sangamon River @ Monticello 77 0.0813 0.0225 0.0636 -0.0096 0.0280 0.1005 0.0595 5576000 South Fork Sangamon R. @ Rochester 42 0.1103 -0.1658 0.1498 0.0732 0.1127 0.1730 -0.1405 5578500 Salt Creek @ Rowell 49 0.0561 0.1188 0.0051 0.0765 0.0544 0.0085 -0.0884 5582000 Salt Creek @ Greenview 50 0.1282 0.1735 0.0645 0.1037 0.0678 0.0955 - 0.0547 5583000 Sangamon River @ Oakford 52 0.1252 0.2392 0.0679 0.1599 0.1342 0.1267 - 0.0211 5584500 LaMoine River @ Colmar 47 0.03420.0300 0.0879 0.0194 -0.0564 -0.0009 - 0.2063 5585000 LaMoine River @ Ripley 70 0.0907 0.0426 0.2108 -0.0385 0.0037 0.0932 - 0.0559 5587000 Macoupin Creek @ Kane 51 0.0196 0.0841 0.0196 0.0667 0.1404 0.0714 - 0.1828

SOUTHWESTERN ILLINOIS 5592000 Kaskaskia River @ Shelbyville 51 0.0620 0.2669 -0.2816 0.0714 0.2047 -0.1184 - 0.0259 5592500 Kaskaskia River @ Vandalia 77 0.0800 0.1081 0.0178 0.0697 0.1408 0.0260 0.0602 5594000 Shoal Creek near Breese 46 0.0647 0.0586 0.1633 0.0454 0.1130 0.0666 - 0.1536 5597000 Big Muddy River @ Plumfield 77 0.0041 0.4323 -0.1709 -0.0171 -0.0123 0.0991 0.1094 @ Plumfield (1915-1969) 55 -0.0626 0.0049 -0.0936 -0.1502 -0.0841 0.0653 - 0.0357 5599500 Big Muddy River @ Murphysboro 61 0.1027 0.5593 0.0525 0.0973 0.1388 0.0710 0.1508

MISSISSIPPI / OHIO RIVERS 5474500 Mississippi River @ Keokuk 113 0.1050 0.2759 0.0972 0.0648 0.2335 0.1092 - 0.0136 7010000 Mississippi River @ St. Louis 55 0.2391 0.4228 0.0896 0.3670 0.3212 0.1838 0.0788 3611500 Ohio River @ Metropolis 57 0.1263 0.3710 -0.0363 0.4499 0.1379 -0.0073 0.1147

Note: Coefficients in bold indicate significance at the 95 percent level of confidence

Table 2. Streamgage Records Used to Analyze Impacts of Climate Variability

Watershed drainage Period of Gage # Location area (mi2) record

03345500 Embarras River at Ste. Marie 1,516 1915- 1991 03346000 North Fork Embarras near Oblong 318 1941- 1991 03379500 Little Wabash River below Clay City 1,131 1915-1991 03380500 Skillet Fork at Wayne City 464 1929-1991 03612000 Cache River at Forman 244 1925-1991 05419000 Apple River near Hanover 247 1935-1991 05430500 Rock River at Afton, WI 3,340 1915-1991 05435500 Pecatonica River at Freeport 1,326 1915- 1991 05438500 Kishwaukee River at Belvidere 538 1941-1991 05440000 Kishwaukee River near Perryville 1,099 1941-1991 05444000 Elkhorn Creek near Penrose 146 1941-1991 05446500 Rock River near Joslin 9,549 1941-1991 05447500 Green River near Geneseo 1,003 1937-1991 05466000 Edwards River near Orion 155 1941-1991 05520500 Kankakee River at Momence 2,294 1916- 1991 05525000 Iroquois River at Iroquois 686 1945-1991 05526000 Iroquois River near Chebanse 2,091 1924- 1991 05527500 Kankakee River near Wilmington 5,150 1916-1991 05542000 Mazon River near Coal City 455 1940-1991 05550000 Fox River at Algonquin 1,403 1915-1991 05555300 Vermilion River Leonore 1,251 1932-1991 05556500 Big Bureau Creek at Princeton 196 1937-1991 05567500 Mackinaw River near Congerville 767 1945- 1991 05569500 Spoon River at London Mills 1,072 1943- 1991 05570000 Spoon River at Seville 1,636 1915-1991 05572000 Sangamon River at Monticello 550 1915-1991 05585000 LaMoine River at Ripley 1,293 1922-1991 05592500 Kaskaskia River at Vandalia 1,940 1915- 1969 05597000 Big Muddy River at Plumfield 794 1915-1969

Table 3. Percentage Increase in Mean Streamflow: 1966-1991 Mean Flow Compared to the 1941-1991 and 1915- 1991 Means

Percent increase in mean flows Region and 1966-1991 1966-1991 gage number Location to 1941-1991 to 1915-1991

Kankakee River vicinity 05520500 Kankakee River at Momence 12.1 20.3 05525000 Iroquois River at Iroquois 11.0 05526000 Iroquois River near Chebanse 13.0 16.8 05527500 Kankakee River near Wilmington 15.5 26.4 05542000 Mazon River near Coal City 16.1

Northern Illinois 05419000 Apple River near Hanover 7.8 05430500 Rock River at Afton, WI 11.0 12.4 05435500 Pecatonica River at Freeport 5.9 6.5 05438500 Kishwaukee River at Belvidere 21.2 05440000 Kishwaukee River near Perryville 20.3 05444000 Elkhorn Creek near Penrose 14.2 05446500 Rock River near Joslin 13.5 05447500 Green River near Geneseo 15.2 05550000 Fox River at Algonquin 19.0 25.0

Central Illinois 05466000 Edwards River near Orion 11.7 05555300 Vermilion River near Leonore 15.0 05556500 Big Bureau Creek at Princeton 15.4 05567500 Mackinaw River near Congerville 12.6 05569500 Spoon River at London Mills 11.6 05570000 Spoon River at Seville 11.3 16.3 05572000 Sangamon River at Monticello 10.7 12.9 05585000 LaMoine River at Ripley 9.8 14.1

Southern Illinois 03345500 Embarras River at Ste. Marie 9.0 9.0 03346000 North Fork Embarras near Oblong 10.2 03379500 Little Wabash River below Clay City 7.5 9.1 03380500 Skillet Fork at Wayne City 4.7 03612000 Cache River at Forman 3.0 2.4 05592500 Kaskaskia River at Vandalia 8.6 12.8 05597000 Big Muddy River at Plumfield 2.3 3.0

Table 4. Influence of Period of Record on Trend Analysis

Kendall coefficient of trend Mean Low High Period of record flow flow flow

Embarras River at Ste. Marie 1915-1991 0.0376 0.0786 0.0595 1915-1964 -0.0743 -0.1616 -0.0122 1924-1973 -0.1118 0.0425 -0.0286 1933-1982 0.0416 0.2738 0.0188 1942-1991 0.0433 0.1395 0.0449

Little Wabash River near Clay City 1915-1991 0.0451 0.0449 0.0581 1915-1964 -0.0824 -0.2024 -0.1102 1924-1973 -0.0514 -0.0408 0.0629 1933-1982 0.0188 0.2891 0.0612 1942-1991 0.0449 0.1990 0.1102

Sangamon River at Monticello 1915-1991 0.0813 0.0225 0.0636 1915-1964 -0.0596 -0.0629 0.0220 1924-1973 -0.0661 0.0255 0.0188 1933-1982 0.1184 0.1939 0.0890 1942-1991 0.0743 -0.0187 0.0612

Spoon River near Seville 1915-1991 0.1237 0.1193 0.0793 1915-1964 -0.0563 -0.1360 0.0384 1924-1973 -0.0122 0.0255 -0.0563 1933-1982 0.1755 0.3708 0.0400 1942-1991 0.0726 0.1582 -0.0057

Note: Bold italics indicate the Kendall coefficient of trend passes the 95 percent level of confidence

Kendall coefficient of trend Mean Low High Period of record flow flow flow

Pecatonica River at Freeport 1915-1991 0.0191 0.1158 -0.1490 1915-1964 -0.1167 -0.2109 -0.0726 1924-1973 -0.0400 -0.0493 0.0024 1933-1982 0.0629 0.2517 0.0824 1942-1991 0.0416 0.2432 -0.1722

Rock River at Afton, WI 1914-1991 0.0849 -0.0032 -0.0056 1915-1964 -0.1886 -0.3350 -0.1526 1924-1973 0.0106 -0.0255 0.0171 1933-1982 0.1820 0.2874 0.0906 1942-1991 0.1951 0.2840 0.0893

Fox River at Algonquin 1916-1991 0.1979 0.2843 0.0758 1916-1964 -0.1599 -0.0957 -0.1667 1924-1973 0.0841 0.1276 0.1200 1933-1982 0.2702 0.4524 0.1069 1942-1991 0.3159 0.4575 0.1510

Kankakee River at Momence 1916-1991 0.3137 0.2252 0.2688 1916-1964 0.0731 0.1879 0.0170 1924-1973 0.1412 0.0289 0.1347 1933-1982 0.3078 0.1582 0.3616 1942-1991 0.2310 0.0476 0.2816

Kankakee River near Wilmington 1916-1991 0.3558 0.2685 0.2618 1916-1964 0.0782 0.1294 0.0714 1924-1973 0.1151 -0.0034 0.0629 1933-1982 0.3012 0.1871 0.1886 1942-1991 0.3176 0.1752 0.1788

Table 5. Number of Occurrences of Significant Low-Flow Events

Region and Decade gaging location 1920s 1930s 1940s 1950s 1960s 1970s 1980s

Kankakee River Vicinity Kankakee River at Momence 3 3 2 2 3 0 1 Iroquois River near Chebanse 2 3 2 1 3 0 1 Kankakee River near Wilmington 3 3 1 3 3 0 1 Mazon River near Coal City -- -- 3 5 1 0 1 Regional Average 2.7 3.0 2.0 2.8 2.5 0.0 1.0

Northern Illinois Apple River near Hanover -- 3 1 3 3 2 1 Rock River at Afton, WI 0 8 1 5 3 1 1 Pecatonica River at Freeport 0 7 2 3 1 0 1 Kishwaukee River at Belvidere -- -- 1 2 3 1 2 Kishwaukee River near Perryville -- -- 3 3 2 2 1 Elkhorn Creek near Penrose -- -- 3 4 1 1 0 Rock River near Joslin -- -- 2 4 2 1 1 Green River near Geneseo -- -- 1 5 1 2 2 Regional Average 0.0 6.0 1.8 3.7 2.0 1.2 1.1

Central Illinois Edwards River near Orion -- -- 1 4 3 2 2 Vermilion River near Leonore -- 3 0 3 2 1 1 Mackinaw River near Congerville -- -- -- 4 2 0 3 Spoon River at London Mills -- -- -- 3 2 1 2 Spoon River at Seville 1 5 1 4 2 1 2 Sangamon River at Monticello 1 3 0 2 2 1 3 LaMoine River at Ripley 1 3 1 2 1 2 3 Regional Average 1.0 3.5 0.4 3.1 2.0 1.1 2.3

Southern Illinois Embarras River at Ste. Marie 2 2 1 2 3 1 1 North Fork Embarras near Oblong -- -- 0 3 3 0 2 Little Wabash River below Clay City 1 3 1 4 2 0 1 Skillet Fork at Wayne City -- 3 2 2 3 0 4 Cache River at Forman 1 3 2 1 2 4 1 Regional Average 1.3 2.8 1.2 2.4 2.6 1.0 1.8

Table 6. Number of Significant Flood Events

Region and Decade gaging location 1920s 1930s 1940s 1950s 1960s 1970s 1980s

Kankakee River Vicinity Kankakee River near Wilmington winter 3 0 6 2 2 2 5 summer 1 1 2 3 1 4 3 Northern Illinois Pecatonica River at Freeport winter 6 4 9 4 3 2 2 summer 0 0 1 1 1 0 1 Central Illinois Spoon River at Seville winter 2 2 5 3 3 4 3 summer 6 1 5 3 2 3 4 Sangamon River at Monticello winter 5 2 4 3 4 5 6 summer 2 1 1 1 2 5 5 LaMoine River at Ripley winter 3 10 5 5 5 5 3 summer 2 2 6 4 3 4 7 Southern Illinois Embarras River at Ste. Marie winter 5 4 8 4 4 6 4 summer 2 1 2 2 2 3 1 Cache River at Forman winter 8 5 11 4 4 8 4 summer 2 3 0 2 1 1 2 Table 7. Number of Significant Flood Days

Region and Decade gaging location 1920s 1930s 1940s 1950s 1960s 1970s 1980s

Kankakee River Vicinity Kankakee River near Wilmington winter 5 0 14 3 6 10 27 summer 0 5 13 8 5 3 6 Northern Illinois Pecatonica River at Freeport winter 24 20 26 30 7 8 9 summer 0 0 3 5 6 0 5 Central Illinois Spoon River at Seville winter 3 5 15 11 4 12 9 summer 18 3 12 6 6 10 16 Sangamon River at Monticello winter 8 5 8 10 9 11 11 summer 5 3 5 2 5 10 8 LaMoine River at Ripley winter 7 22 41 21 17 19 27 summer 6 8 23 10 13 10 22 Southern Illinois Embarras River at Ste. Marie winter 11 12 26 9 8 20 12 summer 6 3 10 8 8 5 4 Cache River at Forman winter 22 15 31 4 8 15 7 summer 5 6 0 8 4 3 7

Table 8. Correlation of Trend Functions to Annual Mean Discharges

Linear Cyclic Step Location trend trend trend

Kankakee River at Momence 0.3296 0.2689 0.3642 Rock River at Afton, WI 0.1070 0.2450 0.3886 Fox River at Algonquin 0.2239 0.3145 0.3754

Table 9. Streamgages Used to Analyze Impacts of Urbanization

Watershed Additional drainage Period of modifications Gage # Location area (mi2) record to watershed

05528500 Buffalo Creek at Wheeling 19.6 1953-1991 SWD 05529500 McDonald Creek near Mt. Prospect 7.9 1953- 1991 SWD 05530000 Weller Creek at DesPlaines 13.2 1951-1991 RF, SWD 05530500 Addison Creek at Bellwood 17.9 1950-1991 RF, SWD 05533000 Flag Creek near Willow Springs 16.5 1952- 1991 RF 05531500 Salt Creek at Western Springs 115.0 1946- 1991 RF, SWD 05536340 Midlothian Creek at Oak Forest 12.6 1951- 1991 SWD 05536500 Tinley Creek near Palos Park 11.2 1952-1991 05537500 Long Run near Lemont 20.9 1952-1991 RF 05550500 Poplar Creek at Elgin 35.2 1952-1991 RF, SWD

Notes: RF = major return flows (wastewater treatment facility discharges) to the stream. SWD = stormwater detention reservoirs affect high flows.

Table 10. Streamgages Used to Analyze Impacts of Major Reservoirs

Watershed drainage Period of Gage # Location area (mi2) record Reservoir

05550000 Fox River at Algonquin 1,403 1915-1991 Fox Chain of Lakes 05573540 Sangamon River at Decatur 938 1982-1991 Lake Decatur 05578500 Salt Creek near Rowell 335 1943-1991 Clinton Lake 05592000 Kaskaskia River at Shelbyville 1,054 1941-1991 Lake Shelbyville 05592500 Kaskaskia River at Vandalia 1,940 1915- 1991 Lake Shelbyville 05593000 Kaskaskia River at Carlyle 2,719 1938- 1991 Carlyle Lake 05597000 Big Muddy River at Plumfield 794 1915-1991 Rend Lake 05599500 Big Muddy River at Murphysboro 2,169 1931-1991 Rend Lake

Table 11. Streamgages Used to Analyze Impacts of Water Use

Watershed drainage Period of Source of Gage # Location area (mi2) record return flows

WITHDRAWALS 05554500 Vermilion River at Pontiac 579 1943-91 05576000 S. Fork Sangamon River near Rochester 867 1950-91

RETURN FLOWS 03339000 Vermilion River near Danville 1,290 1929- 91 Danville 05439500 S. Branch Kishwaukee River near Fairdale 387 1941-91 DeKalb 05469000 Henderson Creek near Oquawka 432 1935-91 Galesburg 05576500 Sangamon River at Riverton 2,618 1915- 56, Decatur 1986-91 05580950 Sugar Creek near Bloomington 34.4 1975-91 Bloomington-Normal 05581500 Sugar Creek near Hartsburg Bloomington-Normal 05595200 Richland Creek near Hecker 129 1970-91 Belleville

RETURN FLOWS - URBAN WATERSHEDS 05530000 Weller Creek at DesPlaines 13.2 1951-91 05533000 Flag Creek near Willow Springs 16.5 1952- 91 Hinsdale 05531500 Salt Creek at Western Springs 115.0 1946- 91 MWRDGC 05537500 Long Run near Lemont 20.9 1951-91 Chickasaw Hills, Derby Meadows 05550500 Poplar Creek at Elgin 35.2 1952-91 Streamwood

Figure 1. Comparison of annual precipitation and streamflow, Embarras River watershed near Charleston.

Figure 2. Location of streamgages and identification of trends in a) mean flows, b) low flows, c) high flows, and d) seasonal flows.

Figure 3. Annual series of mean flow for seven long- term gaging stations. Bar charts show the annual series. Solid lines show 11-year moving averages.

Figure 4. Annual series of low flows for five long-term gaging stations. Bar charts show the annual series. Solid lines show 11-year moving averages.

Figure 5. Annual series of high flows for seven long- term gaging stations. Bar charts show the annual series. Solid lines show 11-year moving averages.

Figure 6. Annual precipitation for Marengo and streamflow for the Kishwaukee River at Belvidere.

Figure 7. Annual precipitation for Kankakee and streamflow for the Kankakee River near Wilmington.

Figure 8. Comparison of the annual mean discharge series to a linear trend, cyclic trend, and step trend on the a) Kankakee River at Momence, b) Fox River at Algonquin, and c) Rock River at Afton, Wisconsin.

Figure 9. Annual peak discharge series for Addison Creek at Bellwood.

Figure 10. Annual peak discharge series for Midlothian Creek at Oak Forest.

Figure 11. Effect of reservoir storage and evaporation on the outflow of Crystal Lake in McHenry County (from Knapp, 1988).

Figure 12. Lake Decatur inflow and outflow.

Figure 13. Annual 7-day low-flow series of the Fox River at Algonquin. Increases since the 1940s are caused by changes in reservoir operation from the Fox Chain of Lakes.

Figure 14. Annual 7-day low-flow series of the Big Muddy River at Plumfield. Increases since 1972 are caused by minimum flow releases from Rend Lake.

Figure 15. Annual 7-day high flow series of the Kaskaskia River at Carlyle. Reductions since 1967 are caused by flood control from Carlyle Lake.

Figure 16. Annual 7-day low-flow series of the South Fork Sangamon River near Rochester. Extreme low flows in 1976, 1988, and 1989 are caused by withdrawals.

Figure 17. Percentage of the 7-day low flow in Illinois streams that originates from return flows.

Return to Table of Contents , Volume 2