Long-Term Changes in Lake Temperature
Text below is based on the 2007 report, Major Lakes Continuous Temperature Study: Interim Progress Report, prepared by King County. Please see the report for additional details and citations to relevant publications.
Long-term changes in the temperature dynamics of Lake Washington have been observed, including a trend of increasing temperatures and related biological effects. The surface 10 meters and whole lake temperatures have increased over the period 1964-1998 by 1.5 oC (0.0458 oC per year) and 0.98 oC (0.0268 oC per year), respectively. The warming trend has been greatest for the whole lake from April to September, for the epilimnion from August to October and for the hypolimnion in March and April.
Regional Warming Influence
This trend toward increased lake temperatures is best attributed to the trend of increasing atmospheric long wave radiation – consistent with expectations from human-induced global climate warming. In addition, there is a significant influence of the Pacific Decadal Oscillation (PDO), and to a lesser extent the El Niño Southern Oscillation (ENSO), on lake temperatures. These changes are not unique to Lake Washington, but have been observed in other large lakes throughout the world (Table 1).
Table 1. Summary of published lake warming trends.
|Lake||Location||Period||Warming Rate, oC per year||Basis h|
|Lake Washingtonb||Seattle, WA||1964-1998||0.026||VWA|
|Lake 239c||NW Ontario||1964-1998||0.108||DA|
|Lake Tanganyikaf||E. Africa||1913-1975||0.0042||DA|
|Lake Tanganyikaf||E. Africa||1975-2000||0.0039||DA|
|Lake Malawi* g||E. Africa||1939-1999||0.01||VWA|
|*(>300 m depth)|
Adapted from Coats et al. (2006)
a Coats et al. (2006)
b Arhonditsis et al. (2004)
c Schindler et al. (1996)
d Ambrossetti and Barbanti (1999)
e Livingstone (2003)
f Verberg et al. (2003)
g Vollmer et al. (2005)
h VWA – Volume-weighted average; DA– Depth average
Temperature, Lake Thermal Stratification, and Biological Responses
Researchers have also shown that long-term warming trends in Lake Washington have resulted in lengthening the period spring-summer stratification period by 25 days over the last 40 years (1962-2002) – mainly due spring stratification beginning about 16 days earlier than it did 40 years ago. Again, the influence of the PDO, and to a lesser extent ENSO, superimposes variability on the long-term trends in stratification onset and termination.
In response to earlier onset of stratification, the spring phytoplankton bloom occurs about 19 days earlier than it did in 1962. Researchers have suggested that this might be a factor in an observed decline in Daphnia abundance (the main consumers of the spring phytoplankton bloom), with implications for food supply to upper trophic levels in the lake. The timing of the spring peaks of two species of Daphnia (D. pulicaria and D. thorata) are also undergoing complex phenological shifts, with implications for sockeye salmon fry specifically and other zooplankton and planktivores in general.
It has also been hypothesized that the progressive warming of Lake Washington in recent years may be leading to more physically stable lake conditions during summer that could provide a competitive advantage to Microcystis, a blue-green alga or cyanobacterium. This blue-green species is capable of producing neurotoxins (nerve) and hepatotoxins (liver) toxins, with implications for the health of humans and other animals exposed directly through recreational activities.
Rising lake temperatures have also been suggested as a possible explanation for the loss of adult sockeye salmon as they migrate through Lake Washington in recent years. Elevated water temperature in general has been identified as a potentially significant factor in the decline of ESA-listed salmon, although this concern is related more specifically to temperature increases in local streams resulting from clearing of shading vegetation, channel and wetland modifications, and water diversions.
Lake Stratification and Thermal Stability
To illustrate the changes in lake stability that have already been documented, the trends in the annual onset, end and duration of stratification and the annual maximum lake stability based on data provided by the University of Washington for the period 1964-2006 are shown in Figure 1. Based on a non-parametric trend test (Mann-Kendall), with the exception of the date the lake becomes destratified, these metrics have statistically significant trends (p<0.05). In general, the lake has become more stable earlier in the year and remains stable longer which translates into a longer duration of lake stability. Annual maximum lake stability is also increasing.
Figure 1. Annual estimates of (A) stratification onset, (B) destratification, (C) duration of stratification, and (D) maximum Schmidt Stability Index based on the University of Washington data for the period 1964-2006.
Note: Stability threshold of 150 g-cm cm-2 was used. All trends were evaluated using the non-parametric Mann-Kendall test.
Index of Thermal Stress on Salmonids
Previous analyses have suggested a relationship between the number of days the temperature exceeds some threshold and Chinook salmon escapement (including harvest) in the Lake Washington system. Elevated water temperatures are also suspected to have caused the apparent loss of approximately half of the 2004 sockeye salmon run between the Ship Canal Locks and sockeye spawning grounds in the Cedar River and other Lake Washington tributaries.
With a chosen lethal threshold, an Index of Thermal Stress (ITS) can be calculated which essentially produces degree-day values above a chosen threshold. A threshold of 20 oC was chosen for use here because it is the upper incipient lethal temperature (UILT) for salmon – the water temperature at which approximately half of the population would survive with permanent exposure. This threshold is also consistent with the observations that during unusually warm years (mean daily temperatures above 20 oC) Fraser River sockeye hormonal and stress indicators suggested that fish were suffering significant physical stress and maturation impairment. Exposure of salmon to temperatures in excess of 20 oC also appears to be associated with a much higher risk of disease.
In general, peak temperatures of 22-23 oC occur in August near the peak migration period (Figure 2). Temperature begins to fall, but are still above the threshold by the end of the sockeye migration through the locks.
Figure 2. Plot of average surface temperature (0 to 5-m depth) observed at 0512 and 0512B between 2004-2006.
Note: 2006 Data for 0512 and 0512B are incomplete. Red vertical lines indicate the range of dates used to calculate the Index of Thermal Stress. Daily adult sockeye counts at the Ballard Locks (black bars) are also shown. These data were obtained from the Washington Department of Fish and Wildlife website.
To evaluate the ITS for potential trends, the cumulative ITS was also calculated for the long term temperature data collected in Lake Washington by the UW for the period 1964 to 2006. These results, along with the results for the routine King County and continuous temperature monitoring location near the locks, are plotted in Figure 3. The long-term data collected in Lake Washington appear to be a reasonable surrogate for thermal stress experienced by salmon migrating through the Ship Canal. This is not surprising as the temperature of the surface of the lake in summer is likely driven almost exclusively by surface heat exchange processes across the lake surface – including the Ship Canal. Although there is no statistically significant trend over the period 1964-2006, there is a distinct upward trend beginning in 1980, with the highest ITS estimates occurring in 1998 and 2004 – 2 years that have anecdotally been noted for high incidence of pre-spawn mortality (Chinook in 1998 and sockeye in 2004). The upward trend in thermal stress is consistent with observed warming trends in Lake Washington – daily maximum temperatures were not necessarily higher (mid-summer daily temperatures typically exceed 20 oC), but periods of elevated temperatures each year appear to be longer and/or more frequent.
Figure 3. Plot of annual total Index of Thermal Stress based on the University of Washington long-term data set, and stations 0512 and 0512B for the period 1964-2006.
For more information about changing temperatures in Lake Washington, please see the 2007 Major Lakes Continuous Temperature Study: Interim Progress Report, prepared by King County.