George B. Arhonditsis

Challenges and Opportunities for Integrating Lake Ecosystem Modelling Approaches

A large number and wide variety of lake ecosystem models have been developed and published during the past four decades. We identify two challenges for making further progress in this field. One such challenge is to avoid developing more models largely following the concept of others (‘reinventing the wheel’). The other challenge is to avoid focusing on only one type of model, while ignoring new and diverse approaches that have become available (‘having tunnel vision’). In this paper, we aim at improving the awareness of existing models and knowledge of concurrent approaches in lake ecosystem modelling, without covering all possible model tools and avenues. First, we present a broad variety of modelling approaches. To illustrate these approaches, we give brief descriptions of rather arbitrarily selected sets of specific models. We deal with static models (steady state and regression models), complex dynamic models (CAEDYM, CE-QUAL-W2, Delft 3D-ECO, LakeMab, LakeWeb, MyLake, PCLake, PROTECH, SALMO), structurally dynamic models and minimal dynamic models. We also discuss a group of approaches that could all be classified as individual based: super-individual models (Piscator, Charisma), physiologically structured models, stage-structured models and trait-based models. We briefly mention genetic algorithms, neural networks, Kalman filters and fuzzy logic. Thereafter, we zoom in, as an in-depth example, on the multi-decadal development and application of the lake ecosystem model PCLake and related models (PCLake Metamodel, Lake Shira Model, IPH-TRIM3D-PCLake). In the discussion, we argue that while the historical development of each approach and model is understandable given its ‘leading principle’, there are many opportunities for combining approaches. We take the point of view that a single ‘right’ approach does not exist and should not be strived for. Instead, multiple modelling approaches, applied concurrently to a given problem, can help develop an integrative view on the functioning of lake ecosystems. We end with a set of specific recommendations that may be of help in the further development of lake ecosystem models.

Multiple regression models: A methodology for evaluating trihalomethane concentrations in drinking water from raw water characteristics

The presence of trihalomethanes (THMs) in drinking water has attracted the attention of both researchers and professionals, because of the harmful effects of these substances on human health. A multiple regression model was developed to estimate THM concentrations in finished drinking water, using data from the Menidi Treatment Plant of Athens. A number of routinely measured characteristics--including chlorine dose, chlorophyll a, temperature, pH and bromide--of raw water, were used to generate a reliable methodology for predicting both total THM and individual species concentrations. Seasonality effects were also considered during the analysis. In general, these models were found to give acceptable fits, estimating accurately lows and highs over the annual cycle.

The effects of seston lipids on zooplankton fatty acid composition in Lake Washington, Washington, USA

We collected suspended particulate matter (seston) and zooplankton samples from Lake Washington in Seattle, Washington, U.S.A., over a 10-month period to investigate the effects of food availability on zooplankton fatty acid (FA) composition. The percentage of nutritionally critical omega3 polyunsaturated fatty acids (PUFA) in the seston varied from 8% of the FA pool in midsummer to 30% during the spring diatom bloom. Zooplankton accumulated much higher percentages omega3 PUFA than was available in the seston. In particular, cladocerans preferentially accumulated eicosapentaenoic acid (EPA, 20:5omega3), copepods accumulated docosahexaenoic acid (DHA, 22:6omega3), and both copepods and cladocerans accumulated 18 carbon chain omega3 PUFAs (C18 omega3). By comparison, the FA of zooplanktivorous juvenile sockeye salmon (Oncorhynchus nerka) were strongly dominated by EPA (12.5% +/- 2.1%) and DHA (28.2% +/- 8.7%). The saturated fatty acid and the arachidonic acid (ARA, 20:4omega6) composition of Diaptomus ashlandi was strongly (r2 = 0.76) and moderately (r2 = 0.54) correlated with the prevalence of these FAs in the seston. Furthermore, the DHA content of Diaptomus was moderately correlated with the sestons DHA content (r2 = 0.45) and very strongly correlated with seston EPA (r2 = 0.89). Since EPA was the most prevalent PUFA in the seston and DHA was the most prevalent PUFA in Diaptomus, these results suggest that Diaptomus may synthesize DHA from the EPA in their food. In general, zooplankton species in Lake Washington were strongly enriched with those FA molecules that are most physiologically important for fish nutrition (i.e., ARA, EPA, and DHA), indicating a clear mechanism by which changes in seston composition influence fisheries ecology.

A community-based framework for aquatic ecosystem models

Here, we communicate a point of departure in the development of aquatic ecosystem models, namely a new community-based framework, which supports an enhanced and transparent union between the collective expertise that exists in the communities of traditional ecologists and model developers. Through a literature survey, we document the growing importance of numerical aquatic ecosystem models while also noting the difficulties, up until now, of the aquatic scientific community to make significant advances in these models during the past two decades. Through a common forum for aquatic ecosystem modellers we aim to (i) advance collaboration within the aquatic ecosystem modelling community, (ii) enable increased use of models for research, policy and ecosystem-based management, (iii) facilitate a collective framework using common (standardised) code to ensure that model development is incremental, (iv) increase the transparency of model structure, assumptions and techniques, (v) achieve a greater understanding of aquatic ecosystem functioning, (vi) increase the reliability of predictions by aquatic ecosystem models, (vii) stimulate model inter-comparisons including differing model approaches, and (viii) avoid ‘re-inventing the wheel’, thus accelerating improvements to aquatic ecosystem models. We intend to achieve this as a community that fosters interactions amongst ecologists and model developers. Further, we outline scientific topics recently articulated by the scientific community, which lend themselves well to being addressed by integrative modelling approaches and serve to motivate the progress and implementation of an open source model framework.

The re-eutrophication of Lake Erie: Harmful algal blooms and hypoxia

Lake Erie supplies drinking water to more than 11 million consumers, processes millions of gallons of wastewater, provides important species habitat and supports a substantial industrial sector, with >

Quantification of the effects of nonpoint nutrient sources to coastal marine eutrophication: applications to a semi-enclosed gulf in the Mediterranean Sea

50 billion annual income to tourism, recreational boating, shipping, fisheries, and other industries. These and other key ecosystem services are currently threatened by an excess supply of nutrients, manifested in particular by increases in the magnitude and extent of harmful planktonic and benthic algal blooms (HABs) and hypoxia. Widespread concern for this important international waterbody has been manifested in a strong focus of scientific and public material on the subject, and commitments for Canada-US remedial actions in recent agreements among Federal, Provincial and State agencies. This review provides a retrospective synthesis of past and current nutrient inputs, impairments by planktonic and benthic HABs and hypoxia, modelling and Best Management Practices in the Lake Erie basin. The results demonstrate that phosphorus reduction is of primary importance, but the effects of climate, nitrogen and other factors should also be considered in the context of adaptive management. Actions to reduce nutrient levels by targeted Best Management Practices will likely need to be tailored for soil types, topography, and farming practices.

Phosphorus retention in a mesotrophic lake under transient loading conditions: insights from a sediment phosphorus binding form study.

An integrated modelling approach is proposed for the assessment of the nutrient loading of a coastal marine ecosystem from terrestrial sources. The evaluation of the model was based on a data set collected on a monthly basis from a Greek gulf surrounded by an intensively cultivated watershed. It consisted of three interacting components, that is the terrestrial, the hydrodynamic and the biological submodels. A modification of the loading-functions approach formed the basis of the terrestrial submodel, used for the estimation of the nutrient fluxes due to agricultural run-off. Nutrient loads from the atmosphere, the domestic effluents and the industrial activity were also considered. The hydrodynamic submodel, based on the Princeton Ocean Model, was used for the estimation of the spatial transport of nutrients and organic carbon. The third component was the biological submodel that was focused on the interactions of nitrogen, phytoplankton, zooplankton, bacteria and organic carbon. The model has shown good fit to the experimental data and it was further used for the assessment of the role of various exogenous and endogenous sources of nutrients and organic carbon. It was found that nutrient loading from the agricultural run-off was remarkable during winter, the contribution being between 40 and 60% of the total nitrogen stock. The fluxes of nitrogen and organic carbon from sewerage and industrial activity were also quantified and they were estimated to be up to 10% of the total stock whereas, the contribution of the wet and dry deposition from the atmosphere was insignificant. The integrated modelling approach could also be used for the understanding of the energy flow through the marine food web and the testing of various scenaria concerning the development of the coastal zone.

Predicting the Frequency of Water Quality Standard Violations Using Bayesian Calibration of Eutrophication Models

Phosphorus retention in sediments has been estimated for three basins in Lake Simcoe, a mesotrophic lake in Ontario, Canada. Total phosphorous (TP) fractionation was used to examine the concentration of phosphorus (P) binding forms in the sediments of Cooks Bay, Kempenfelt Bay, and the Main Basin. The extended sequential extractions allowed us to differentiate between organic-, inorganic-, carbonate-bounded and redox-sensitive phosphorus. Our results showed different mechanisms of P release in each of the three investigated basins, which may be linked to their distinct loading histories, present land-uses and morphology of the sampling sites. In the deep Main Basin, where moderate changes in P loading have been induced by deforestation, sediments are not an important long-term source of diagenetically mobile P, as almost 75% of P is released within a short time scale. P release is predominantly generated by a continuous epilimnetic P flux, rather than a large inventory of temporary P stored in the sediments. Diagenesis in the upper sediment layers is fast enough to prevent a large accumulation of temporary P. In the much deeper glacially formed Kempenfelt Bay with a highly urbanized catchment, P release from the sediments is dominated by the redox-sensitive P fraction, representing up to 40% and 57% of long- and short-term sediment P release, respectively. In the shallow and agriculturally-impacted Cooks Bay, the main P binding form that can be mobilized through diagenesis is carbonate-bound P. This fraction contributes 40.1% and 37.6% to the long- and short-term P sediment release, respectively. Although different mechanisms of P release have been revealed for the three basins in Lake Simcoe, the vertical profiles indicate that the sediments throughout the system are still able to bind deposited P.

Mass Flux Calculations Show Strong Allochthonous Support of Freshwater Zooplankton Production Is Unlikely

The water quality standard setting process usually relies on mathematical models with strong mechanistic basis, as this provides assurance that the model will more realistically project the effects of alternative management schemes. From an operational standpoint, the interpretation of model results should be coupled with rigorous error analysis and explicit consideration of the predictive uncertainty and natural variability. In this study, our main objective is to attain effective model calibration and rigorous uncertainty assessment by integrating environmental mathematical modeling with Bayesian analysis. We use a complex aquatic biogeochemical model that simulates multiple elemental cycles (org. C, N, P, Si, O), multiple functional phytoplankton (diatoms, green algae and cyanobacteria) and zooplankton (copepods and cladocerans) groups. The Bayesian calibration framework is illustrated using three synthetic datasets that represent oligo-, meso- and eutrophic lake conditions. Scientific knowledge, expert judgment, and observational data were used to formulate prior probability distributions and characterize the uncertainty pertaining to a subset of the model parameters, i.e., a vector comprising the 35 most influential parameters based on an earlier sensitivity analysis of the model. Our study also underscores the lack of perfect simulators of natural system dynamics using a statistical formulation that explicitly accounts for the discrepancy between mathematical models and environmental systems. The model reproduces the key epilimnetic temporal patterns and provides realistic estimates of predictive uncertainty for water quality variables of environmental management interest. Our analysis also demonstrates how the Bayesian parameter estimation can be used for assessing the exceedance frequency and confidence of compliance of different water quality criteria. The proposed methodological framework can be very useful in the policy-making process and can facilitate environmental management decisions in the Laurentian Great Lakes region.

The effects of episodic rainfall events to the dynamics of coastal marine ecosystems: applications to a semi-enclosed gulf in the Meditteranean Sea

Many studies have concluded terrestrial carbon inputs contribute 20–70% of the carbon supporting zooplankton and fish production in lakes. Conversely, it is also known that terrestrial carbon inputs are of very low nutritional quality and phytoplankton are strongly preferentially utilized by zooplankton. Because of its low quality, substantial terrestrial support of zooplankton production in lakes is only conceivable when terrigenous organic matter inputs are much larger than algal production. We conducted a quantitative analysis of terrestrial carbon mass influx and algal primary production estimates for oligo/mesotrophic lakes (i.e., TP≤20 µg L−1). In keeping with the principle of mass conservation, only the flux of terrestrial carbon retained within lakes can be utilized by zooplankton. Our field data compilation showed the median (inter-quartile range) terrestrial particulate organic carbon (t-POC), available dissolved organic carbon (t-DOC) inputs, and in-lake bacterial and algal production were 11 (8–17), 34 (11–78), 74 (37–165), and 253 (115–546) mg C m−2 d−1, respectively. Despite the widespread view that terrestrial inputs dominate the carbon flux of many lakes, our analysis indicates algal production is a factor 4–7 greater than the available flux of allochthonous basal resources in low productivity lakes. Lakes with high loading of t-DOC also have high hydraulic flushing rates. Because t-DOC is processed, i.e., mineralized or lost to the sediments, in lakes at ≈0.1% d−1, in systems with the highest t-DOC inputs (i.e., 1000 mg m−2 d−1) a median of 98% of the t-DOC flux is advected and therefore is not available to support zooplankton production. Further, advection is the primary fate of t-DOC in lakes with hydraulic retention times <3 years. When taking into account the availability and quality of terrestrial and autochthonous fluxes, this analysis indicates ≈95–99% of aquatic herbivore production is supported by in-lake primary production.