Search

This document is a corrigendum that corrects an error in a previously published paper titled "Monitoring tropical freshwater fish with underwater videography and deep learning." The authors apologize for the mistake, which involves the incorrect color of the bounding box and species name for Ambassis macleayi in Figure 7e. The correct color should have been yellow instead of green. The authors provide a corrected version of the figure and apologize for any confusion caused. [Extracted from the article]

Context: The application of deep learning to monitor tropical freshwater fish assemblages and detect potential anthropogenic impacts is poorly understood. Aims: This study aimed to compare the results between trained human observers and deep learning, using the fish monitoring program for impact detection at Ranger Uranium Mine as a case study. Methods: Fish abundance (MaxN) was measured by trained observers and deep learning. Microsoft's Azure Custom Vision was used to annotate, label and train deep learning models with fish imagery. PERMANOVA was used to compare method, year and billabong. Key results: Deep learning model training on 23 fish taxa resulted in mean average precision, precision and recall of 83.6, 81.3 and 89.1%, respectively. PERMANOVA revealed significant differences between the two methods, but no significant interaction was observed in method, billabong and year. Conclusions: These results suggest that the distribution of fish taxa and their relative abundances determined by deep learning and trained observers reflect similar changes between control and exposed billabongs over a 3-year period. Implications: The implications of these method-related differences should be carefully considered in the context of impact detection, and further research is required to more accurately characterise small-growing schooling fish species, which were found to contribute significantly to the observed differences. Tropical freshwater fish assemblages monitored in Kakadu National Park were measured using underwater videography, and counted by trained human observers and artificial intelligence (AI). Over 3 years of monitoring, both AI and trained human observers detected the same changes in overall fish assemblage composition. However, AI consistenly underestimated the abundance of small-growing schooling fish. Image credit: Andrew Jansen. This article belongs to the collection Ecological Monitoring and Assessment of Freshwater Ecosystems: New Trends and Future Challenges. [ABSTRACT FROM AUTHOR]

Riparian trees located in seasonally dry environments may be reliant on groundwater supplies, but the prevalence and magnitude of groundwater uptake is often unclear. Using soil water matric potential and water stable isotopes, we examined the relative contributions of soil water and groundwater to the dry season water uptake of five riparian tree species along an intermittent river of tropical northern Australia. Because xylem water was depleted in deuterium relative to source water (average offset −14.0‰), we numerically removed this offset and assessed the effect of the correction on mixing model results. We also estimated the isotopic composition of unbound soil water (i.e., the portion of soil water not tightly bound to soil particles) from bulk soil water data by using an empirical formulation from the literature and tested whether considering unbound soil water as a source would affect our results. Despite the hot and dry surface environment, we found that soil moisture was available for trees at relatively shallow (~0.7–1.5 m) depths. When unbound soil water and corrected xylem water data were considered, most tree species used a combination of this soil moisture source and groundwater from the capillary fringe. However, not correcting for isotopic effects resulted in large underestimations of the groundwater contributions to tree water uptake. Our findings suggest that ignoring soil isotopic effects and deuterium depletion in xylem water may reduce the validity of source water partitioning assessments. Further research is needed on the likely causes for deuterium depletion in xylem water. [ABSTRACT FROM AUTHOR]

Magnesium (Mg) is a primary contaminant in mine water discharges from the Ranger Uranium Mine (north Australia). Site‐specific water quality guideline values (WQGVs) for Mg have been derived from laboratory and field studies. Contaminated groundwater with elevated electrical conductivity and metals (Mg, Mn, U, SO4, and Ca) was detected flowing from the mine site into adjacent surface waters. This provided an opportunity to investigate the protectiveness of the Mg WQGV by conducting an integrated laboratory and field study. A direct toxicity assessment (DTA) of the groundwater was conducted with local tropical freshwater species: duckweed (Lemna aequinoctialis), green hydra (Hydra viridissima), and the aquatic snail Amerianna cumingi. An in situ toxicity assessment was carried out in the creek receiving diluted groundwater by use of the same species of snail, to aid interpretation of laboratory‐derived data. The toxicity of the contaminated groundwater was higher than Mg‐only toxicity testing for H. viridissima, with other elevated metals and major ions contributing to toxicity. However, for duckweed and snail, the contaminated groundwater was less toxic than the Mg‐only testing. In situ snail monitoring supported laboratory exposures, showing no effect on reproduction of A. cumingi exposed to an average of approximately 5 mg/L Mg; however, a very small effect was noted closer to the groundwater source, probably associated with other contaminants. The minimal toxicity observed for L. aequinoctialis and A. cumingi, despite the elevated Mg, can be explained by the high calcium (Ca) concentration of the water and the potential amelioration of metal toxicity. The extent of Ca amelioration of Mg toxicity was organism dependent. This study affirms the proposed environmental rehabilitation standard of 3 mg/L Mg for surface waters with a Ca concentration typical of water from this mine site. Integr Environ Assess Manag 2019;15:64–76. © 2018 SETAC Key Points: The toxicity of a contaminated groundwater from a uranium mine in north Australia was assessed with a suite of tropical freshwater species.An integrated laboratory and field study was conducted to investigate the protectiveness of the water quality guideline value for 1 of the key potential contaminants of concern, magnesium.Results show that keeping chronic exposure to Mg is kept below 2.5–5 mg/L Mg should protect aquatic ecosystems downstream of the mine.The results support the proposed environmental rehabilitation standard for the mine of 3 mg/L Mg for surface waters of a 3.5:1 Mg:Ca ratio. [ABSTRACT FROM AUTHOR]

The chronic toxicity of ammonia to tropical freshwater species is understudied, and thus data on temperate species have been used to derive water quality guideline values for tropical regions. Such practices may lead to underprotective guideline values due to differences in toxicities observed between tropical and temperate species. In addition, the presence of ammonia in low‐ionic‐strength waters may also result in higher toxicity, and studies on this factor are limited. The present study assessed the toxicity of ammonia to 6 tropical freshwater species in low‐ionic‐strength waters. Because ammonia toxicity varies depending on the pH and temperature, test water pH concentrations were maintained at approximately pH 6.0 ± 0.3 at temperatures between 27.5 and 30 °C. Low‐effect chronic inhibition concentrations were derived for the following species: Chlorella sp. 66 mg L−1; Lemna aequinoctialis 22 mg L−1; Hydra viridissima 1.8 mg L−1; Moinodaphnia macleayi 27 mg L−1; Amerianna cumingi 17 mg L−1; and Mogurnda mogurnda 5.4 mg L−1 total ammonia nitrogen. Two of the species tested (a cnidarian and a fish species) were among the most sensitive reported anywhere within their taxonomic group. Chronic ammonia datasets representing toxicity estimates for temperate and tropical species were plotted and compared using species sensitivity distributions. The results indicate that the differences in chronic toxicity observed between tropical and temperate species were likely due to the low ionic strength of the waters to which tropical species were exposed, rather than any inherent physiological differences between species from tropical and temperate regions. This finding suggests that tropical waters of low ionic strength may be at a higher risk from ammonia compared with other freshwater ecosystems. Environ Toxicol Chem 2019;38:177–189. © 2018 Commonwealth of Australia. Published by Wiley Periodicals, Inc. on behalf of SETAC. [ABSTRACT FROM AUTHOR]