Evolution of a river management industry in Australia reveals meandering pathway to 2030 UN goals

The paper investigates whether Australia’s stream management industry is evolving in line with UN Sustainable Development Goals (SDGs) and the UN Decade on Ecosystem Restoration. Over recent decades, river management in many western nations has shifted from engineering-centric approaches to practices that protect and restore ecological, social and cultural values. In this context, the authors analyze 25 years (1996–2021) of Australian Stream Management (ASM) conference proceedings to understand who participates in the industry, what work is undertaken, and how it is conducted. They test hypotheses aligned with the UN 2030 agenda: (1) Who: the industry is integrating more diverse perspectives, as evidenced by collaboration patterns, institutional diversity, and gender. (2) What: the work has matured in focus (greater nature-based approaches, inclusion of First Nations and community values) and become more interdisciplinary and transdisciplinary. (3) How: the industry is transitioning from reactive to strategic management in response to natural events and is increasing policy influence and adaptive management. The purpose is to document industry evolution, identify gaps, and recommend actions to meet global goals for integrated water resources management and ecosystem restoration.

The authors situate their study within global initiatives (UN MDGs, SDGs, and the Decade on Ecosystem Restoration) that call for accelerated, nature-based, collaborative water resources management. Prior reviews have assessed the success of river repair and restoration projects globally, largely focusing on on-ground outcomes and scientific quality. Australian stream management has historical roots in integrated catchment management and community involvement (e.g., Waterwatch, Landcare). The literature also highlights the need for diverse perspectives, inclusion of Indigenous knowledge, transdisciplinarity, and adaptive practices to build resilience to climate change and natural events. International comparisons (North America, Europe, New Zealand) show growing collaboration and private sector roles, but persistent gaps in community and First Nations integration and challenges in policy impact and adaptive management uptake. Unlike parts of Europe, increases in nature-based solutions are not evident over time in Australia, though some techniques have become standard practice.

Data source and scope: The study analyzes the ASM conference proceedings (1996–2021), which include 958 peer-reviewed six-to-eight-page papers and some abstracts. The dataset reflects a broad cross-section of industry practitioners (academics, government, consultants, utilities, NGOs/community). Structured review: All 958 items were reviewed to catalog author demographics (gender, institution type and location), collaboration patterns (cross-institutional, interstate, international), and to flag content documenting adaptive management (success/failure) and policy (advances/failures). Nine annotators divided work to reduce inter-annotator variability. Trends were tested via linear regression against year with Benjamini–Hochberg correction across 78 tests, yielding a significance threshold p<0.01 (0.01<p<0.05 interpreted as weak evidence). Collaboration networks were visualized for 1996 and 2021 with nodes for nine institution types and edges for co-authorship proportions. Delegate data: Delegate lists for three conferences (1ASM, 9ASM, 10ASM) were digitized to compare attendee vs author representation by gender, affiliation type, and location. Topic modeling: Full-text body was extracted from PDFs using GROBID; abstracts under 300 words (n=26) were excluded, leaving 932 documents. Standard preprocessing removed punctuation, special characters, digits, and stopwords; uncommon words (<5 documents) were excluded; no stemming/lemmatization was applied. A Structural Topic Model (STM) was fitted (R ‘stm’ v1.3.6) across K=10–100 with year as a prevalence covariate; K=60 was selected based on held-out likelihood, semantic coherence/exclusivity, and interpretability. Topics were manually labeled and grouped into disciplines (abiotic science, biotic science, integrative science, community participation, management/policy) and flagged for adaptive management or policy relevance. Interdisciplinarity: a paper was cross-disciplinary if prevalence >0.1 for integrative topics or >0.1 for both an abiotic and a biotic topic; transdisciplinary if prevalence >0.1 for a science topic and >0.1 for a community participation or management/policy topic. Topic-year coefficients (unstandardized b) and p-values identified trends (p<0.01 significant; 0.01<p<0.05 weak). Topic associations were assessed via cosine similarity among top-20 topics per conference year. Bias assessment considered author-specific vocabulary and temporal shifts; labels were validated by multiple experts. Reproducibility: Data and R code are available at https://osf.io/ybkur/.

Who (diversity and collaboration): • Authorship team sizes increased over 25 years, indicating greater collaboration. • Proportion of women authors rose from 18% to 37%, though women remain a minority. • Cross-institution collaboration increased; interstate collaboration remained <21% and international collaboration <8%. • Consultant authorship increased; government/utility and local/federal employees increased their representation as authors but decreased as first authors; state government first authorship declined from 25% (1996) to 11% (2021). • NGO/community authorship remained low throughout. • By 2021, collaboration networks were broader/stronger across nearly all institution types; strong links persisted between universities and consultants, with growing consultant–state government collaboration. Catchment groups, utilities, and local governments increased in representation but were less connected as co-authors. What (focus and interdisciplinarity): • Topic diversity increased. Strong positive trends included: waterway management programs; water and mining; environmental water planning and evaluation. Indigenous river management showed a weak positive trend but remained low in prevalence. • Community participation topics (Indigenous management, eDNA/citizen science, landholder participation, community programs) remained steady overall. • Technology-driven topics (remote sensing; eDNA/citizen science) increased. • Nature-based approaches were already embedded by 1996 and held steady (e.g., geomorphic character and recovery; riparian management; rehabilitation planning). Engineering topics (bed/bank protection structures) remained low or showed no increasing trend; waterway design showed a weak positive trend. • Cross-disciplinary papers (biotic+abiotic) peaked in 2001–2004 and stabilized at ~30–40%; transdisciplinary papers (science + policy/community) fluctuated around ~30–40%, with early peaks (1999–2004), dip (2007), and later increase. • In 1996, strong topic associations were mainly within domains (e.g., abiotic); by 2021, associations bridged science, policy/management, and community, indicating greater transdisciplinarity. How (responses, policy, adaptive management): • Natural events shaped discourse: drought (Millennium drought) aligned with peaks in environmental flow science (early 2000s), followed by water reform (2007), then environmental watering and regulated rivers (2012), and later environmental water planning and evaluation (2021), suggesting progression from science to policy to evaluation. • Flood hydrology peaked in 2012 after major La Niña floods (2010–2012); flood risk management became dominant from 2012 through 2021, reflecting sustained management focus and iterative strategy development after subsequent floods. • Vegetation fire response topics peaked post-major fire events (2003; 2008–2009; 2019–2020), with stronger engagement after repeated post-fire flood/erosion impacts. • Adaptive management topics increased to 1999, peaked 2004–2007, then remained steady; early 2000s emphasis on monitoring/evaluation/reporting was not sustained. Vision-setting, prioritization, and adjustment phases were weakly represented. • Documentation of rehabilitation success increased, but reporting of failures remained rare; policy advances were discussed more than policy failures, suggesting positivity bias and incomplete adaptive management culture. Global context: • Patterns of collaboration and consultant involvement mirror international trends; cross-/transdisciplinarity rates (~35%) appear higher than in the strictly academic river-research literature. • Persistent gaps include low integration of First Nations and community perspectives and limited growth in nature-based solutions compared to some regions (e.g., Europe). • Policy influence and active adaptive management remain limited, similar to challenges reported in the US and Europe.

The study’s findings largely support the hypotheses that Australia’s stream management industry has matured in collaboration, diversity, and interdisciplinarity (Who and What). Collaboration networks now integrate researchers, consultants, and government more fully, and topic associations increasingly span scientific and policy/community domains. However, critical gaps persist: NGO/community and First Nations participation remain low; on-ground community participation has not increased measurably; and adoption of fully proactive adaptive management has stalled, with limited reporting of failures and weak representation of vision-setting/prioritization. Responses to natural events show a maturing sequence from science generation to policy formulation and evaluation (e.g., environmental flows; flood risk), but this has not generalized into a pervasive, learning-oriented adaptive management cycle. The limited growth in nature-based solutions contrasts with EU trends, suggesting a need for renewed emphasis and scaling. Collectively, these outcomes indicate partial alignment with UN SDG targets (6.5, 6.6) but reveal structural and cultural barriers—insufficient community/First Nations integration, episodic resourcing, and weak policy influence—that must be addressed to accelerate progress toward 2030.

This 25-year longitudinal analysis of Australia’s stream management industry reveals substantial maturation in collaboration, diversity, and interdisciplinarity, but limited increases in community participation, incomplete adoption of proactive adaptive management, and modest policy influence. To help meet UN SDG targets for 2030 (implement integrated water resources management at all levels and protect/restore water-related ecosystems), the authors recommend: 1) Work holistically to create inclusive communities of practice that integrate diverse stakeholders, especially First Nations knowledge and participation. 2) Implement and scale nature-based solutions, moving beyond command-and-control engineering and avoiding cycles of unsustainable maintenance. 3) Improve stability and resourcing to enable active, hypothesis-driven adaptive management, including open repositories and cultural change to share failures. 4) Preserve and use institutional and community knowledge more effectively amid institutional change, strengthening governance and collaboration continuity. 5) Step-change industry influence on policy and deliver large-scale stream/catchment management through ethical, just, and transparent decision-making and stronger two-way engagement with policy-makers. These actions can help translate maturing practice into transformative progress aligned with the UN Decade on Ecosystem Restoration.

Findings derive from ASM proceedings—a knowledge-sharing forum representing a subset of the Australian stream management industry, predominantly from south-eastern states. Geographic biases include under-representation of Western Australia, South Australia, and the Northern Territory, and likely under-representation of Aboriginal and Torres Strait Islander peoples. Conference themes and locations may subtly influence topics and attendance distributions, though long-term trends appear robust. Proceedings emphasize pragmatic applications over theoretical advances; some purely academic developments may be under-captured, while practice-oriented insights are likely overrepresented. Topic modeling may be influenced by changing vocabulary and author-specific language; expert validation of topic labels mitigated but did not eliminate such biases. Collaboration co-authorship does not necessarily capture all real-world collaborations (e.g., localized entities may participate in practice without co-authoring). Overall, these factors may limit generalizability across Australia and internationally, though results are likely most representative of south-eastern Australia and comparable developed contexts.