Final Report: Eutrophication Thresholds – Assessment, Mitigation, and Resilience in Landscapes and Lakes

EPA Grant Number: R832445
Title: Eutrophication Thresholds – Assessment, Mitigation, and Resilience in Landscapes and Lakes
Investigators: Carpenter, Stephen R. , Foley, Jonathan A. , Turner, Monica G.
Institution: University of Wisconsin Madison
EPA Project Officer: Packard, Benjamin H
Project Period: September 1, 2005 through June 30, 2007 (Extended to August 31, 2008)
Project Amount: $299,999
RFA: Exploratory Research: Understanding Ecological Thresholds In Aquatic Systems Through Retrospective Analysis (2004) RFA Text |  Recipients Lists
Research Category: Aquatic Ecosystems , Water


Eutrophication, a persistent environmental problem characterized by turbid water, toxic algae, fish kills, waterborne disease, and loss of aquatic ecosystem services, may be related to important thresholds in the phosphorus (P) cycle. We will address two main questions: (1) What thresholds in the transport and recycling of P in linked terrestrial-aquatic ecosystems cause lakes to switch between clear-water and eutrophic states? (2) How can thresholds for transport and recycling of P be manipulated to mitigate eutrophication, or increase resilience of clear water lakes against eutrophication? The research will conduct a retrospective analysis of the Yahara watershed and its major lakes (near Madison, Wisconsin), using a substantial historical data base of land characteristics and limnology.

Summary/Accomplishments (Outputs/Outcomes):

We defined thresholds as switchpoints in the response of an environmental variable to a controlling variable, such that the behavior of the environmental variable on one side of the threshold cannot be simply extrapolated from the behavior on the other side of the threshold. A simple example is a light switch, where the environmental variable is the state of the lamp and the controlling variable is the state of the switch. Most environmental thresholds are more complicated. For example, soils may take up dissolved phosphorus from groundwater, or release it to groundwater, depending on the bulk concentration of phosphorus in the soil and several other controlling variables related to soil chemistry.

A subset of environmental thresholds are critical thresholds. In critical thresholds, the rate equations that govern the time evolution of environmental variables approach different kinds of equilibria, depending on thresholds in controlling variables. The oldest and most famous example in aquatic ecology is the paradox of enrichment, where algae approach a stable equilibrium if nutrient inputs are low, and approach a cyclic equilibrium when nutrient inputs are high.

We reviewed thresholds in watershed phosphorus (P) cycles that might be important in eutrophication. We identified the following thresholds:

P application and soil buildup: fertilizer application to soil may increase soil P concentration or not, depending on plant uptake

Bank and slope stability: as the intensity of storms or duration of storms increases, for a slope of given gradient, a threshold is crossed where the soil slumps downslope

Drainage density and gully formation: as drainage area and slope increase, a threshold is crossed where rainstorms of a certain intensity cause gullies to form.

P sorption and desorption: above a certain level of soil P, for a soil of given chemical characteristics, any additional phosphorus added to the soil will be soluble.

Sediment detachment or deposition: above a certain intensity of rainfall, soil particles are detached; below a certain intensity of rainfall, soil particles are deposited.

Runoff generation: above a threshold of water content, for a given soil, precipitation results in overland flow.

Watershed runoff generation increases sharply at thresholds in precipitation intensity.

A simulation model of the Lake Mendota watershed was developed to study the responses of these thresholds to plausible scenarios of climate change and land use change. It was not possible to complete publishable reports of these results prior to the end of funding. Dr. Daniel Collins, the postdoc on the project, has moved to a position in New Zealand. He states that he will publish the results from his new position.

We contributed to a synthesis of how ecological thresholds may be useful for environmental management, participating in an EPA-sponsored workshop and development of a manuscript (Groffman et al. 2006). In this review, the methods for identifying and investigating thresholds are discussed with a variety of terrestrial and aquatic environments across a range of spatial scales.

We also studied a critical threshold in the P cycle of lakes. Beyond a certain level of P input (the eutrophication threshold), recycling of P from sediments occurs at a rate that will maintain eutrophication even if P input rate is decreased. P input rate must be decreased to a second threshold, the oligotrophy threshold, which is lower than the eutrophication threshold. This phenomenon has been studied for a long time. Several different chemical and biological mechanisms can be involved, and several models have been used to describe this behavior of lakes. We extended this work in two ways.

First, we showed that several statistical properties of time series for water P change in characteristic ways before a eutrophication threshold is crossed (Carpenter and Brock 2006). We focused on increased variance because of the availability of robust estimators for variance as a function of time scale in ecological time series. We showed that simple time-series filters can amplify the signal and increase prospects for anticipating thresholds of eutrophication. A manager could use these signals to decrease P input rates before the threshold was crossed.

In a second paper about the leading indicators, we considered a stylized model of ecosystem services subject to critical transitions and connected to an economy that depends on the ecosystem for both disposal of waste products and generation of a consumable ecosystem service such as freshwater or clean air (Brock and Carpenter 2006). We showed that the indicators perform consistently in this setting, that spatial patterns of the indicators are also sensitive to incipient crossing of thresholds, that simple multivariate decompositions can separate signals of impending transitions from “junk” variability in the environment, and that economic data can provide advance warning of ecological thresholds based in the divergence of prices “put” and “call” options in futures markets for ecosystem services. These results suggest applications for management of ecosystem services that are spatially heterogeneous and are marketed, such as freshwater in some parts of the world.

In our second study of the critical threshold for eutrophication, we estimated the threshold for Lake Mendota, Wisconsin (Carpenter and Lathrop 2008). Our goal was to estimate the probability of crossing thresholds for water quality of the lake. We used 30 years of P budget data as well as independent measurements of some key rates. Data were used in a well-studied model of lake phosphorus dynamics to compute probability distributions of thresholds using Markov-chain Monte Carlo simulation.

Results were surprising in three main ways. (1) The probability that future changes in the lake’s phosphorus would be smooth and continuous, with no thresholds, was only 3.4%. Thus future changes in the lake are likely to involve thresholds. (2) Probability distributions for thresholds were wide. It was impossible to assign a precise value to the thresholds. (3) Thresholds for oligotrophy and eutrophy were rather close together, indicating that phosphorus trends could switch direction over rather small ranges of phosphorus input. If phosphorus inputs are decreased below an oligotrophy threshold, then the lake moves toward a low-phosphorus clear-water state. If phosphorus inputs are increased above a threshold for eutrophy, then the lake moves toward a high-phosphorus turbid state.

The findings support current management efforts to decrease phosphorus inputs to Lake Mendota. Even though the lake has demonstrated fast improvements after past droughts, there is no guarantee that this resilience will continue forever. On the contrary, if phosphorus inputs increase then the lake could cross a eutrophy threshold, leading to a long-lasting period of poor water quality. On the other hand, if phosphorus inputs are decreased there is a chance for crossing an oligotrophy threshold to a long-lasting period of good water quality. This would be a windfall for managers and the public. Precise values of the thresholds cannot be determined. Despite the uncertainty of the threshold values, the advice to managers is clear – reduce phosphorus inputs to avoid the eutrophy threshold and increase the chance of crossing the oligotrophy threshold.

For scientists, the wide probability distributions of the thresholds are sobering. It is not possible to precisely define the numerical values for the thresholds, even though the database for Lake Mendota is among the best in the world. Simulations show that we could measure the thresholds precisely if they are crossed (Carpenter, S.R., 2003, Regime Shifts in Lake Ecosystems, Ecology Institute, Oldendorf/Luhe, Germany). However, crossing the eutrophy threshold would be an expensive disaster. Phosphorus management may some day take the lake across the oligotrophy threshold, and if that occurs then long-term data might allow us to measure the threshold value rather precisely.

Our study shows that ecosystem thresholds for eutrophication are hard to measure precisely, even with extensive data. Nonetheless, we show how probability distributions for ecosystem thresholds can be quantified accurately and inclusively. Fortunately, this analysis leads to clear policy advice for eutrophication mitigation.

Journal Articles on this Report : 4 Displayed | Download in RIS Format

Other project views: All 8 publications 4 publications in selected types All 4 journal articles
Type Citation Project Document Sources
Journal Article Brock WA, Carpenter SR. Variance as a leading indicator of regime shift in ecosystem services. Ecology and Society 2006;11(2):9. R832445 (2006)
R832445 (2007)
R832445 (Final)
  • Full-text: Ecology and Society
  • Abstract: Ecology and Society
  • Other: Ecology and Society PDF
  • Journal Article Carpenter SR, Brock WA. Rising variance: a leading indicator of ecological transition. Ecology Letters 2006;9(3):311-318. R832445 (2006)
    R832445 (2007)
    R832445 (Final)
  • Abstract from PubMed
  • Journal Article Carpenter SR, Lathrop RC. Probabilistic estimate of a threshold for eutrophication. Ecosystems 2008;11(4):601-613. R832445 (Final)
  • Abstract: Ecosytems
  • Journal Article Groffman PM, Baron JS, Blett T, Gold AJ, Goodman I, Gunderson LH, Levinson BM, Palmer MA, Paerl HW, Peterson GD, Poff NL, Rejeski DW, Reynolds JF, Turner MG, Weathers KC, Wiens J. Ecological thresholds: the key to successful environmental management or an important concept with no practical application? Ecosystems 2006;9(1):1-13. R832445 (Final)
    R828012 (Final)
    R828677C001 (Final)
    R832441 (Final)
  • Full-text: UWisc-PDF
  • Abstract: Springer-Abstract
  • Other: ResearchGate-Abstract & Full Text-PDF
  • Supplemental Keywords:

    soil, sediments, risk, health effects, ecological effects, bioavailability, vulnerability, susceptibility, cumulative effects, chemicals, toxics, pathogens, indicators, restoration, scaling, Bayesian, ecology, physics, environmental chemistry, modeling, Midwest, Wisconsin, WI, EPA Region 5, RFA, Scientific Discipline, Air, Ecosystem Protection/Environmental Exposure & Risk, Aquatic Ecosystems & Estuarine Research, climate change, Air Pollution Effects, Aquatic Ecosystem, Environmental Monitoring, Ecological Risk Assessment, Atmosphere, anthropogenic stress, estuarine research, eutrophication, landscape change, ecological thresholds, anthropogenic impact, ecosystem indicators, modeling ecosystem change, lakes, aquatic ecosystems, water quality, ecosystem stress, trophic interactions, ecosystem response, environmental history

    Relevant Websites: Exit

    Progress and Final Reports:

    Original Abstract
  • 2006 Progress Report
  • 2007 Progress Report