The temperature of spring and stream water within the Spring Creek watershed in central Pennsylvania is important for many water uses, especially maintaining healthy aquatic ecosystems, water recreation and a highly-valued trout fishery. Trout are stressed when water temperatures exceed about 68.0 degrees Fahrenheit (°F) and temperatures above about 74.2 °F can be lethal. Many headwater mountain streams and springs emerge from the forested uplands with non-carbonate rocks such as sandstone and shale and relatively stable temperature regimes, which fluctuate around the regional mean annual air temperature of about 50.2 °F. Waters then flow downstream into the karst valleys underlain by carbonate limestone and dolomite rocks where they mix with other springs and streams to form the main stem Spring Creek and its major tributaries: Buffalo Run, Big Hollow, Slab Cabin Run, Thompson Run, Cedar Run, and Logan Branch (see Figures 1 and 2). As these streams flow across and down the valleys, various types of land uses can affect their temperature in a mixed land-use watershed typical of the Spring Creek basin, especially in urban growth centers of State College, Penn State campus and Bellefonte.
Figure 1. Map of Spring Creek Watershed showing stream temperature monitoring sites.
Figure 2. Map of Spring Creek Watershed showing locations of spring water temperature monitoring sites.
Out of concern over impacts of rapid urban-suburban development on water resources, the Spring Creek Watershed Association initiated a water monitoring program in 1997 with broad community support. The project included quarterly assessment of water quality including temperatures in springs and streams as well as continuous monitoring of stream temperatures at selected stream stations. The program expanded to include point measurements of water temperature on a quarterly basis at 8 spring sites and 14 stream sites as well as continuous temperature monitoring at 18 stream sites. The program continues till this day with data collection and data base management being conducted by the Water Resources Monitoring Project led by a volunteer oversight committee. Laboratory analysis of water chemical and biological quality is provided by the Pennsylvania Department of Environmental Protection. Descriptions of monitoring sites are given in Table 1. A more complete description of the watershed monitoring project, annual data reports, and data access requests can be found here. This paper is based on analysis of available monitoring data generally for 1999-2017.
Table 1. Description of Stream and Spring Temperature Measurement Sites in Spring Creek Watershed
Factors Affecting Stream Temperatures
Stream temperatures are naturally strongly affected by climate as water seeks an equilibrium with air temperatures and variations in solar radiation. An example of the annual pattern of daily maximum and minimum water temperatures during a four-year period at the Spring Creek Park (SPP-see Table 1 for code and description of each station) is shown in Figure 1. Annual maximums generally occur in June through August within the range of about 68 to 73 °F at this site. Annual minimums approach 32 °F each winter during January through February.
Day to day variations in maxima and minima reflect response to local weather conditions including variations in flows due to rainfall and snow-melt runoff. On average, maximum daily temperatures exceed daily minimums by about 5-6 °F at this site, but this daily difference between maximums and minimums can exceed about 12-13 °F on the hottest, low-flow summer days and drop below a few degrees °F on winter days with low sun angles or during high flows. Each stream site follows this same general pattern annually and daily but precise timing and magnitude of temperature fluctuations vary with natural site conditions, especially location of any spring-flow inputs and impacts of watershed land uses.
Stream temperatures are also sensitive to many factors related to land development especially changes in the amount of shading and depths and volumes of water in the channel. Land clearing for urban/suburban/rural housing and commercial developments, farming, roads and interstate highway construction, forest management, utility corridors, etc., can all influence water temperatures. Small dams in channels can also slow and increase surface areas of flowing waters and cause heating in summer. Dam removals can help reduce maximum stream temperatures. Temperatures are often also affected by storm water runoff during rains and snow-melt runoff events. Higher turbidity or cloudiness in storm waters can also increase water temperatures by affecting the amount of solar radiation that is absorbed by stream water.
Discharges of waters into the streams from limestone quarries, sewage treatment plants, fish hatcheries, industry, and natural springs within the watershed can also play a role in controlling stream temperatures depending on the volumes and temperatures of water added. Conversely, major withdrawals of water due to groundwater pumping for water supplies can reduce flows and water depth and influence the temperature regime. All human-caused variations are superimposed on the natural seasonal and daily temperature fluctuations as shown in Figure 3. Consequently, the temperature in Spring Creek at a point is difficult to understand since it reflects a mixture of upstream and local influences of many natural and human-induced factors.
Figure 3. Example of typical daily maximum and minimum stream temperatures found at Spring Creek Park (SPP) during 2013-2017
Spring and Stream Temperature Comparisons
Measurements of spring and stream temperatures at the time of seasonal or quarterly water sampling provided unique comparisons of temperatures between springs and streams on the same days. Figure 4 shows the average, maximum and minimum water temperatures at the time of sampling over the period of record (4-18+ years) for all springs and all stream sites sampled. The overall averages for springs (50.5 °F) and streams (51.8 °F) sites were similar to each other since over years the daily and seasonal fluctuations tend to average out. These temperatures are also similar to the average annual air temperature in State College (50.2 °F) since water temperatures tend to approach equilibrium with air temperatures. Stream temperatures generally do lag behind air temperature by a day or two, but the time lags for spring water coming from underground can be much greater. Maximum water temperatures for all springs averaged about 9.7 °F cooler than for stream sites and minimum water temperatures for springs averaged about 7.0 °F warmer than stream sites. This comparison suggests that inputs of spring water into stream channels can have a moderating effect on summer and winter temperatures in streams. The exact impact of spring water on stream temperatures depends upon the location of the spring inputs compared to the stream measurement point and the temperature and volume of spring water added.
Figure 4. Comparison of average stream and spring water temperatures measured at the times of quarterly sampling during 2005-2017 for all data, maximum data and minimum temperature data.
When looking at temperatures of individual springs, you find that not all springs are created equal! Figure 5 shows that maximum and minimum temperatures in all eight springs that were monitored vary considerably. Springs such as Blue Spring (BLS), Walnut Spring (WAS), and Windy Hill Spring (WIS) showed surprisingly variable temperatures, mainly higher maximums, compared to springs such as Linden Hall (LIS) and Axemann Springs (AXS) which have the most moderate temperature regimes. Big Spring in Bellefonte (BIS), Continental Courts Spring (COS), and Benner Spring (BES) showed intermediate temperature fluctuations. Since natural temperature fluctuations in soil and rock material that spring water travels through tend to decrease with depth, the exact path that spring water follows, the time of travel and the nature of rock it contacts can all influence temperatures of emerging spring water. In the karst terrain found in the valleys of Spring Creek watershed, the presence of sinkholes can also allow rapid mixing of surface water with spring water flowing underground and complicate interpretation of spring water temperatures.
Figure 5. Maximum and minimum temperatures measured in springs during quarterly sampling showed considerable variations during 2005-2017.
Water Temperatures and Dissolved Oxygen.
Water temperatures are also important because they control the maximum amount of oxygen that can be dissolved (DO) in water for use by aquatic organisms. The saturation levels of dissolved oxygen (DO) increase as water temperatures decrease. Consequently, streams and lakes which experience increasing temperatures can undergo reduced dissolved oxygen levels, which can lead to fish kills. In Spring Creek watershed streams, dissolved oxygen levels were measured with calibrated, hand-held oxygen meters at the time of seasonal water sampling. For the period of record, DO in all streams averaged 11.5 ppm, which equaled slightly over (103%) of the maximum saturation levels predicted for the measured stream temperatures. These DO levels indicate that Spring Creek stream water was well oxygenated and should support healthy aquatic ecosystems.
Interestingly, measured dissolved oxygen in springs was lower than that for streams, even though springs were often cooler than streams. DO in spring water only averaged 8.3 ppm, or 73% of saturation for average spring temperatures. This seeming contradiction is due to reduced contact by spring water with the atmosphere when traveling underground, while stream water has direct contact with the atmosphere especially in fast-flowing turbulent reaches. Also, in streams algae and rooted aquatic plants produce oxygen during daylight hours which becomes dissolved in the water. Fortunately, even reduced levels of DO in spring water are sufficient for most aquatic organisms to survive and spring water rapidly re-oxygenates after flowing a short distance over the surface with contact with the atmosphere.
Spatial Patterns in Maximum Daily Water Temperatures
Data from continuously-monitored stream sites can be used to help understand how water temperature in the Spring Creek main channel varies as it flows downstream and mixes with the various major tributaries. Patterns in maximum daily water temperatures in Spring Creek and at the mouths of main tributaries can be seen in Figure 6 at continuously-monitored sites from headwaters to mouth of the basin. For reference purposes the daily water temperature data are summarized in Appendix A showing averages of the daily mean, daily maximum, and daily minimum temperatures for all stream sites during their respective periods of record.
Figure 6. Downstream variations in mean daily maximum stream temperatures in Spring Creek and at mouths of major tributaries.
Oak Hall to Spring Creek Park. Maximum daily temperatures in Spring Creek increased from an average of 51.8 °F at the uppermost continuously-monitored site in Oak Hall (SPU) to an average of 54.0 °F at Spring Creek Park (SPP) about 3.1 mi (5 km) downstream. This increase in temperature is likely related to warmer inputs from the Cedar Run basin (CEL) with an average maximum daily temperature of 55.4 °F caused by loss of shade due the dominant agricultural land use along this section. Other possible influences on stream temperature in this section are limestone quarry and urban/suburban/commercial developments as the stream flows along the eastern edge of State College and through Lemont on its way to Spring Creek Park. An unmonitored flow from Thornton Spring also enters in this section near the intersection of College Ave. and Houserville Road. Finally, wherever streams are affected by inputs of cool spring water, some natural warming can occur when air temperatures are warmer than the stream.
Spring Creek Park to Houserville. Below Spring Creek Park maximum daily temperatures remain relatively stable at an average of 54.0 and 54.1 °F, respectively, at the Penn State Sheep Farm (SPS) and Houserville (SPH) sites, even though warmer inputs are received by the stream from Millbrook Marsh (MIL) just below Puddintown Road (Figure 6). Total distance between Spring Creek Park and Houserville sites is only about 1.2 miles (2 km). The Millbrook Marsh tributary input with a mean daily maximum temperature of 55.4 °F represents a mix of water from Slab Cabin Run (SLL) at 56.5 °F, Thompson Run (THL) at 55.4 °F and springs which flow through and from developed area of State College/Penn State. Lack of a temperature response in Spring Creek to the generally warmer Millbrook Marsh inputs is likely due to low flow rates in Millbrook discharge at the time of maximum daily temperatures each summer. In contrast, at times of storm flows, the Thompson Run flows can become a complex mix of water from storm sewers from Easterly Parkway, College Avenue and Penn State sub-drainages, Walnut Spring, Thompson Spring and “Duck Pond” outflow and discharges. Slab Cabin Run flows come from agricultural lands below Pine Grove Mills past State College well fields into suburbs of State College, eventually mixing with Thompson Run in Millbrook Marsh along with additional unmonitored flows from Bathgate Springs in the Marsh. Millbrook Marsh plays a special role in Spring Creek hydrology and water quality control by mixing natural spring flows from unmonitored Bathgate Spring with flows from Thompson Run and Slab Cabin Run.
Houserville to Axemann Gauge. After flowing about 5.0 mi (8 km) below Houserville, Spring Creek reaches its highest average maximum daily temperature of 55.8 °F at the Axemann Gauge (SPA) site. Despite relatively heavy shading in the Spring Creek Canyon portion of this stream reach, the stream exhibits increases in maximum daily temperatures of 1.6 °F on average and on the hottest summer days up to 3.6 °F increase in maximum daily temperatures between Houserville and Axemann Gauge. This warming could be natural or be induced by the many human impacts known to occur between the Houserville and Axemann Gauge stations. These include loss of shade especially around housing and bridge sites, inputs of UAJA treated sewage, stream dams on former prison lands, and discharges from both Benner Springs and Fisherman’s Paradise fish hatcheries. One monitored spring, Benner Spring (BES), does contribute cool water directly to the stream in this section which would help counteract warming. Regardless, the cause of this warming between Houserville and Axemann Gauge, either natural or human-caused, is not fully understood, but needs to be determined to help protect the aquatic ecosystem in this heavily-fished section of Spring Creek. It should be noted that the Axemann Gauge site on Spring Creek was named for the closest community to this USGS stream gauging site, which is Axemann, but the actual community of Axemann and Axemann Spring (to be mentioned later) are actually located over a divide in the Logan Branch sub-drainage.
Axemann Gauge to Milesburg. Between the Axemann Gauge (SPA) and Milesburg (SPM) sites the stream exhibits cooling as it flows through the lower segment of the watershed past Bellefonte (Figure 6). Within this 3.7 mi (6 km) long zone Spring Creek receives relatively cool (mean maximum daily temperature 52.3 °F) water at relatively high flows rates from Logan Branch (LOL) which when mixed with cool flows from Big Spring (BIS) in Bellefonte helps moderate the warmer water temperatures in Buffalo Run (BUL) which has an average maximum daily temperature of 54.1 ° F. The maximum and average temperatures observed in Big Spring flows based upon quarterly sampling during 2005-2017 were 52.3°F and 50.4°F, respectively, which helped cool Spring Creek water. The lower Logan Branch site (LOL) has flows augmented by Axemann Spring (AXS) and possible discharges from a limestone quarry. Spring Creek in this lower section is relatively wide and naturally less shaded which would lead to greater warming by the sun than in upper reaches. Buffalo Run is also largely agricultural in its upper reaches and also quite exposed as it flows through unshaded channel in parts of Bellefonte to the BUL site. Spring Creek does receive discharges from the Bellefonte sewage treatment plant below the community, but the resulting average maximum daily temperature of 55.0 °F at the Milesburg (SPM) is still quite tolerable for trout.
Frequency of Water Temperatures Stressful for Trout
Examination of the frequency of maximum daily water temperatures at the stream monitoring stations that are stressful for trout is another way to evaluate possible impacts of temperature regimes and watershed developments on Spring Creek aquatic ecosystems. The time series of daily maximum water temperatures can be analyzed to compute the percentage of days when temperatures exceed levels which could cause stress (>68 °F) or lethal (>74.2 °F) conditions for trout. For example, maximum daily temperature data for the Spring Creek Park station (SPP) over the approximately 4-year time series in Figure 7 show that stressful temperatures were recorded about 3% of the time and lethal temperatures were never recorded. Available periods of records for all sites varied from about 4 years to over 18 years of data. Frequencies of stressful conditions did not appear sensitive to the length of period of record available. Comparisons of such percentages were made for all monitoring stations to determine where stressful conditions for trout occurred.
Figure 7. Example of determining the percentage of days with maximum temperatures which could cause stressful conditions (> 68 °F) and lethal conditions (>74.2 °F) conditions for trout during 2013-2017 at Spring Creek Park (SPP).
Figure 8 shows the percentage of water temperature stress days for trout for all stations. Stations with 10% or more stressful days for trout occurred at the main Spring Creek site at Axemann Gauge (SPA) site and at four Slab Cabin Run sites (SLU, SL8, SLC, SLL). Lowest occurrence of stressful water temperature conditions of <1% of stress days occurred at Spring Creek at Oak Hall (SPU), Logan Branch at Bellefonte (LOL), and Spring Creek at Milesburg (SPM). High frequency of stress days at the Axemann Gauge site (SPA) corresponds with the high maximum daily temperatures found at that site. High stress days at the Slab Cabin Run sites are in a reach (just upstream and downstream of the Atherton Street crossing) where extremely low stream flow rates are commonly observed during summer low flows, which leads to higher maximum temperatures (see Appendix Table A). Low flows in this Slab Cabin Run reach may be related to high groundwater pumping rates for the State College water supply during summer dry periods. Slab Cabin Run at the SLU site also has already experienced heating due to flow through a long reach of open agricultural lands between Pine Grove Mills and State College. Penn State has recently been conducting additional temperature monitoring on Slab Cabin Run in cooperation with Ferguson Township and the State College Borough Water Authority. High stress days do not necessarily indicate lethal conditions for trout. Trout can seek cooler micro-habitats within a reach, but may not be able to complete their entire life cycle within such reaches.
Figure 8. Comparison of the percentage of days with stressful water temperatures for trout during the period of record for Spring Creek monitoring stations.
Days with lethal temperatures for trout were uncommon at most sites. Three Slab Cabin Run sites (SL8, SLU, and SLC) showed 4-5% of days had lethal temperatures due to extremely low flows during dry periods and flow through open agricultural lands. Walnut Run Middle site (WAM) which is affected by intermittent storm drainage from Easterly Parkway urban/suburban lands in State College also had lethal temperatures on 1.5 % of days likely due to extreme low or intermittent flows. Other sites showed <1% lethal days, with Spring Creek Park (SPP) showing 0% lethal days (see Figure 7).
Long-term Trends in Water Temperature
With the concern over global climate change, it is natural to ask whether Spring Creek water temperatures are changing over the long-term. Trends in maximum and minimum daily water temperatures at stations like the Spring Creek at Axemann Gauge site (SPA) with one of the longest periods of record (18+ years) should be sensitive to climate change. Figure 9 shows that both maximum and minimum daily water temperatures during the period are relatively stable with a slight tendency of both to decrease rather than increase. The slight decrease in water temperatures may be tracking a minor decrease in air temperatures, an increase in precipitation and/or streamflow rates or some other natural or human-caused factor within the watersheds. Facts reported in the 2009 Spring Creek Watershed Monitoring Project Annual Report (published in 2011) suggests that low flows have been increasing during 1944-2009 which could account for the cooling trend. Temperature data for other stream sites with long-term records also suggested stable to slightly negative temperature trends.
Figure 9. Long-term trends in maximum (upper) and minimum (lower) daily stream temperatures at the Axemann Gauge (SPA) site on Spring Creek showing a stable to slightly declining temperature trend.
Stream temperatures within the Spring Creek Watershed show differences in temperature regimes between springs and streams and among stream sites within the drainage network. Spring water was cooler in summer and warmer in winter than stream water. Based upon quarterly sampling during 2005-2017, maximum temperatures in springs averaged 9.7 °F cooler than in streams and minimum spring temperatures averaged 7.0 °F warmer than in streams. Dissolved oxygen levels in streams were generally at 100% of saturation levels for the given stream temperatures, while spring water was at about 83% of saturation due to lack of contact of spring water with the atmosphere at the point of emergence. Regardless, dissolved oxygen levels in both streams and emerging spring waters were sufficient to support healthy aquatic ecosystems and spring water would likely quickly re-oxygenate once at the surface. Maximum daily stream temperature at stream sites could be greatly influenced by the amount and temperature of spring water inputs to the channel upstream of the monitoring site. Reduced stream shading and reduced depth of stream water due to groundwater pumping also appeared to increase the maximum daily stream temperatures at some sites. More data on the temperatures and flow rates of unmonitored water inputs to Spring Creek are needed to fully explain patterns of water temperatures observed with the monitoring network. The frequency of days with maximum daily temperatures that exceeded stress levels for trout of 68 °F varied among stream reaches from 0-5% of stress days for sites with a major to moderate spring inputs, 5-10% of stress days for sites with some loss of stream shade and >10% of stress days for sites with a combination of loss of shade and shallow flows during the hotter summer periods. Stream temperatures at sites with up to 18+ years of monitoring data showed a stable or slight cooling trend over the long-term.
Appendix A. Daily Stream Temperatures (°F) in Spring Creek Watershed (up to 24-May-17)
Dr. DeWalle is retired from teaching and research at Penn State in the Department of Ecosystem Science and Management. His research focused on impacts of climate change, land use change and acid rain on water quality and hydrology of watersheds.