Progress in understanding the vulnerability of freshwater ecosystems

Abstract

The ability to collect and synthesize long-term environmental monitoring data is essential for the effective management of freshwater ecosystems. Progress has been made in assessment and monitoring approaches that have integrated routine monitoring programs into more holistic watershed-scale vulnerability assessments. While the concept of vulnerability assessment is well-defined for ecosystems, complementary and sometimes competing concepts of adaptive management, ecological integrity, and ecological condition complicate the communication of results to a broader audience. Here, we identify progress in freshwater assessments that can contribute to the identification and communication of freshwater vulnerability. We review novel methods that address common challenges associated with: 1) a lack of baseline information, 2) variability associated with a spatial context, and 3) the taxonomic sufficiency of biological indicators used to make inferences about ecological conditions. Innovation in methods and communication are discussed as a means to highlight meaningful cost-effective results that target policy towards heuristic ecosystem-management.

Introduction

Long-term environmental monitoring data is essential for impact evaluation and causal inferences ¹ to plan effective freshwater ecosystem management. With increased development, water managers, policymakers, industry, government(s), and local community groups, have taken stewardship roles in monitoring water quality. ² The challenge in managing freshwater ecosystems is an inherent lack of commonly accepted methods or strategies to ascertain impairment and communicate risk. Water quality monitoring programs will often make decisions based on logistics, cost, capacity, and existing policy and practice; rarely are factors behind the decision-making process adequately studied prior to the commencement of monitoring programs. Yet, awareness of sustainability in developing urban centers has increased in parallel to the demand for ecosystem services, which has highlighted the need for effective monitoring solutions.

We are beginning to move away from baseline assessment programs for the sake of monitoring and moving towards a more integrated watershed management system. Indeed, whereas previous ‘freshwater management’ would focus solely on one ecosystem, government(s) have increasingly moved towards integrated watershed management. ³ Here, we define the concept of vulnerability in freshwater assessment and review challenges associated with the detection and communication of ecological impairment. Progress in understanding freshwater vulnerability is framed by a review of approaches that address: 1) a lack of baseline information, 2) spatial variability, 3) and inferences made from biological indicators. Innovation in the communication of results has also become a prominent factor. Indeed, whereas ‘monitoring’ has historically been targeted as a means for baseline assessment, recent studies have suggested that it occurs within a framework of policy implementation and/or development informed by impact evaluation. ¹

Defining vulnerability in freshwater assessment

The term vulnerability alone is one that is generally understood, but there are many contexts in which it is used, and many complex factors that are difficult to quantify that can inherently influence the potential for loss. ⁴ Vulnerability as a concept is not new in the social dimension, where it is readily applied at a community level to understand the influence of change. ⁵ The use of the term vulnerability in freshwater management is relatively new ⁶ but recognized as a means to conceptualize and communicate risk. ⁷ While there are many approaches to defining vulnerability from an ecosystem perspective, ⁸ a capacity to adapt implies that balance (e.g. resilience) and trajectory are the two principal factors that can define vulnerability. ⁹ Ecological stress challenges this paradigm, and the concept of ecological integrity attempts to capture resilience with respect to stress and disturbance.^10,11 Thus, ecological impairment is a result of some form of stress, which increases the vulnerability associated with reduced adaptive capacity. We suggest that freshwater assessment should include implementation components to integrate methods of baseline monitoring and concepts of vulnerability and risk assessment with more focus on informing policy and management at the watershed scale (Figure 1).

Figure 1.

A review of the progress in freshwater vulnerability assessment in terms of moving from baseline monitoring to management and policy implementation. Implementation components are indicated in green bubbles.

A scan of the literature indexed in SCOPUS with the boolean search teams “Vulnerability AND assessment AND freshwater” found 408 papers published since January 9, 2023. Of these, 232 were deemed to be relevant articles covering the topic of freshwater assessment after a title and abstract scan. The number of publications that include the concept of vulnerability applied to freshwater assessments has increased in recent years, more than doubling since 2009 (Figure 2(a)). While most publications included the concept of vulnerability in freshwater assessment with a broad scope, a large portion of studies focused on the specific topic areas of groundwater, fish conservation, and water security (Figure 2(b)). The large increase in the number of publications that include vulnerability as a concept in freshwater assessment over the last decade highlights the move towards research at broader scales that include socio-ecological concepts.

Figure 2.

The number of publications that include the concept of vulnerability in freshwater assessment; (a) publications by year, and (b) a reflection of the topic area of publications, with general assessment grouped as ‘freshwater assessment’.

Methods of freshwater assessment

An ecosystem-based vulnerability assessment process aims to quantify impairment of the biotic and abiotic environment and identify the extent of threats that would require management intervention. ¹² Broad inclusion of the concept of vulnerability in specific conservation strategies has been applied to fish,^13–15 invasive species,^16–18 and endangered populations.^19,20 However, the characterization of ecological properties, such as water quality, biological attributes, and habitat, is required for site-specific risk assessment. ⁶

Metrics that synthesize relationships between biological indicators and environmental and/or stress gradients have become common in routine monitoring programs. ²¹ For example, the Index of Benthic Invertebrates (IBI), ²² Rapid Bioassessment Protocol (RBP), ¹¹ Australian River Assessment System (AusRivAS), ²³ and AQEM/STAR, ²⁴ are used to assess freshwater and brackish ecosystems globally.^25–28 Morphologically identified biota, such as benthic macroinvertebrates, are commonly used ²⁹ ; however, recent studies have incorporated the use of diatoms, ³⁰ dragonfly larvae, ³¹ fish (and their traits),^32,33 and rare species. ⁷

While multimetric indices have been used for decades to simplify complex individual species-environment relationships to infer environmental conditions, shortcomings still exist. The first, and most prominent, problem is a lack of baseline information for comparisons of past environments or the calculation of a natural trajectory. ¹ Secondly, spatial and temporal variabilities of biota can influence structural changes, causing “noise” in the assessment, which can prevent universal monitoring approaches. ³⁴ Finally, cost-effectiveness can limit the amount of data collected, and the depth to which it is analyzed. A vulnerability assessment approach can also be applied to quantify the functional aspects of ecosystems, whereby aquatic trophic systems are examined with the use of trait-based metrics.^35,36 While the optima and tolerance of biological indicators across stress gradients are used for assessment, species-specific ecological information, such as functional end-points, ¹ is often lacking, and species-environment correlations are assumed to be independent of time. ³⁷

Baseline context

There is often an assumption that historic data will provide an appropriate reference condition for management purposes, yet ecosystems evolve over time on a natural trajectory influenced by multiple collinear, and often competing, natural factors. ³⁸ The natural condition of ecosystems could pre-date human settlement by several centuries, and monitoring programs are usually deployed after a problem is either identified or assumed. ³⁹ Seldom do local governments have the capacity to monitor freshwater systems using a sampling resolution high enough to understand anthropogenic influences from natural variability. ⁴⁰

The ability of paleolimnology to provide decades to centuries of missing data through the analyses of biological indicators preserved in sediment profiles has recently been highlighted as a means to complement the assessment of vulnerability.^41,42 The community composition of indicators in lacustrine sediment can be used to reconstruct past conditions. By filling gaps in monitoring records and quantifying natural variability, paleolimnology provides a means to predict the future trajectory of lake and wetland ecosystems.^43,44 However, Duarte et al. (2009) ⁴⁵ note the disconnect between the expectation that ecosystems would return to a pre-disturbance state following intervention and the reality that ecological trajectories are influenced across time by a multitude of factors that may be independent of the original degradation pathway that restoration efforts were intended for. Regardless, reconstructions of baseline conditions can facilitate freshwater management so that decision-makers can adapt and respond to changes that are both anticipated and unanticipated.

Spatially-sensitive metrics

The use of trait-based methods for freshwater assessment is relatively new, with a large focus on streams and rivers. ⁴⁶ Trait-based metrics are less prone to spatial variability on a regional scale^47,48 and not influenced by biogeographical constraints, unlike taxonomic responses. ⁴⁹ Indeed, the quantification of functional traits, especially feeding traits, can help assess the resilience of ecosystems as these attributes are fairly consistent over time. ⁷ While functional diversity is often the most widely used metric, especially in the assessment of populations of macroinvertebrates ⁵⁰ and fish,^51,52 some studies have found differences in phylogenetic diversity associated with human influence. ⁵³ However, the interplay between local environment, habitat, and spatial processes can prevent “pure” species sorting of indicator taxa. ⁵⁴ This has led to a bias of studies that focus on ecosystem-functioning relationships at a local scale. ⁵⁰

Recent studies have deployed a metacommunity approach, which includes spatial processes and biological interactions as factors; anthropogenic impairment can also influence community structuring and consequently introduce variability to biotic metrics.^34,55 Milošević et al. (2022) ³⁴ suggest that a multifactor approach, which covers multiple sources of variability in aquatic biota, can significantly increase the sensitivity of monitoring programs, lowering “noise” in the models caused by natural variability. Thus, the inclusion of spatially-sensitive metrics can increase our understanding of impairment at a regional scale.

Emergent methods to address taxonomic sufficiency

A traditional morphological approach in biomonitoring attempts to have “as high as possible” taxonomic resolution for inferences to be made based on known environmental relationships with specific taxa. ⁵⁶ However, the identification of species based solely on morphological features can be problematic; the number of misidentifications increases with taxonomic resolution. ⁵⁷ Morphological identification of freshwater biota is also an extremely time-consuming high-expertise-driven process, which may not always be cost-effective or practical for routine monitoring ⁵⁸ ; fewer taxonomists are getting stretched across more fields of inquiry than their capacity can accommodate.^59,60

Historically, the solution to questions of taxonomic sufficiency has focused on metrics that could use lower taxonomic resolution but provide similar environmental inferences. The development of new methods of identification, such as DNA barcoding, could fundamentally change how we monitor and assess vulnerable ecosystems. High-throughput sequencing of environmental samples ⁶¹ can overcome constraints associated with morphological identification. Analysis of hundreds of specimens simultaneously offers an alternative to traditional morphotaxonomic identification, and can usually generate species-level outputs. ⁶² However, meta-barcoding approaches also have challenges, such as unknown threshold values for species delamination, a lack of abundance data that assessments are based on, as well as infrastructure and cost constraints. ⁶³ Although, improvements in technology will likely lead to the expansion of genetic methods in the near future.

Computer-assisted automated deep learning approaches that use computer vision can be used for species identification to improve the cost-effectiveness of biomonitoring approaches. ⁶⁴ Convolutional Neural Networks (CNN) have been applied to a wide variety of animal taxa (zooplankton, macroinvertebrates, corals, fishes, insects, and even mammals). ⁶⁵ However, deep-learning models are sensitive to the amount of training data available, which can be limited in systems with a large number of rare taxa. ⁶⁶ Computer-assisted deep-learning approaches are promising since the accuracy level in species identification done by artificial intelligence is high at a time when human expertise is limited.

Communication of vulnerability

When reviewing progress in understanding the vulnerability of freshwater ecosystems, one must also consider how vulnerabilities are communicated. Current methods of assessment can synthesize complex biological relationships into a metric score, but these are often difficult to communicate for policy and decision-making purposes. Disaster management (e.g. flood forecasting) has long used a probability-based calculation of likelihood to communicate risk, expressed as a recurrence interval. ⁶⁷ The broader public's understanding of recurrence intervals is problematic ⁶⁸ ; yet, a function of risk assessment relies on the underlying ability to base predictions of hazards on the probability they will occur based on past context. ⁶⁷

Risk is defined and communicated in the hazard risk vulnerability assessment (HRVA) process as the likelihood an event will occur magnified by the consequences of the event occurring. ⁶⁹ While developed for emergency management, ⁷⁰ HRVA has recently been used for climate change adaptation and mitigation planning ⁷¹ and freshwater assessment. ⁷² HRVA is based on communication of the severity of disturbances and incorporates socio-economic and physical-environmental priorities of a particular location. Since HRVA was designed for synthesizing risk for broader public discourse, it offers a process in which adaptation and mitigation measures can be aligned to the output to help inform planning practices. While studies of water security have focused on the communication of vulnerability, we feel as though more effort needs to be made to integrate freshwater assessment with a broader implementation strategy (e.g. Figure 1), which could be informed by an assessment metric focused on communication of risk (similar to the HRVA process).

Conclusion

While decision-makers often a lack long-term modeling and forecasting of freshwater vulnerability to predict the trajectory of ecosystems, progress has been made in assessment methods to address current shortcomings in baseline information, spatial variability, and taxonomic sufficiency of indicators used for inferences of ecological condition. Our review shows that freshwater assessments that include the concept of vulnerability have increased in recent years; yet, few focus on the communication of associated risk to a broader public audience. As such, we recommend that the concept of vulnerability should be applied to freshwater assessment and focus on methods that further our ability to produce and communicate meaningful results. Further innovation in the communication of risk will allow decision-makers to target policy toward heuristic ecosystem management.

Author biographies

AS Medeiros is an interdisciplinary researcher who focuses on understanding the influence of environmental stress on freshwater services, past, present, and future. He is particularly motivated in applying new methodologies to community-based research and ecological vulnerability assessment. By combining principles of freshwater ecology, paleolimnology, and risk assessment, he is able to bridge the gap between the natural evolution of freshwater ecosystems and human-induced change that influences ecosystem services. This includes investigation of water security through the lens of sustainability and conservation, municipal planning, and engineering for freshwater supply services.

D Milošević is an expert taxonomist who specializes in the identification of non-biting midges (also known as chironomids). His research aims to refine taxonomic resolution through deep learning approaches to optimize our understanding of species-environment relationships and has pioneered methods in the automated identification of specimens for biomonitoring of European lakes and streams.