Geology topics

Yang Wang

Publications and source records attributed to Yang Wang.

3 recordsLinked to original sources

USGS2024

Fair graph learning using constraint-aware priority adjustment and graph masking in river networks

Accurate prediction of water quality and quantity is crucial for sustainable development and human well-being. However, existing data-driven methods often suffer from spatial biases in model performance due to heterogeneous data, limited observations, and noisy sensor data. To overcome these challenges, we propose Fair-Graph, a novel graph-based recurrent neural network that leverages interrelated knowledge from multiple rivers to predict water flow and temperature within large-scale stream networks. Additionally, we introduce node-specific graph masks for information aggregation and adaptation to enhance prediction over heterogeneous river segments. To reduce performance disparities across river segments, we introduce a centralized coordination strategy that adjusts training priorities for segments. We evaluate the prediction of water temperature within the Delaware River Basin, and the prediction of streamflow using simulated data from U.S. National Water Model in the Houston River network. The results showcase improvements in predictive performance and highlight the proposed model's ability to maintain spatial fairness over different river segments.

Delaware River basin↗

USGS2018

Isotope niche dimension and trophic overlap between bigheaded carps and native filter-feeding fish in the lower Missouri River, USA

Stable carbon and nitrogen isotope values (δ 13 C and δ 15 N) were used to evaluate trophic niche overlap between two filter-feeding fishes (known together as bigheaded carp) native to China, silver carp ( Hypophthalmichthys molitrix ) and bighead carp ( Hypophthalmichthys nobilis ), and three native filter-feeding fish including bigmouth buffalo ( Ictiobus cyprinellus ), gizzard shad ( Dorosoma cepedianum ) and paddlefish ( Polyodon spathula ) in the lower Missouri River, USA, using the Bayesian Stable Isotope in R statistics. Results indicate that except for bigmouth buffalo, all species displayed similar trophic niche size and trophic diversity. Bigmouth buffalo occupied a small trophic niche and had the greatest trophic overlap with silver carp (93.6%) and bighead carp (94.1%) followed by gizzard shad (91.0%). Paddlefish had a trophic niche which relied on some resources different from those used by other species, and therefore had the lowest trophic overlap with bigheaded carp and other two native fish. The trophic overlap by bigheaded carp onto native fish was typically stronger than the reverse effects from native fish. Average niche overlap between silver carp and native species was as high as 71%, greater than niche overlap between bighead carp and native fish (64%). Our findings indicate that bigheaded carps are a potential threat to a diverse and stable native fish community.

Lower Missouri River↗

USGS2017

Integrated species distribution models: combining presence-background data and site-occupancy data with imperfect detection

Two main sources of data for species distribution models (SDMs) are site-occupancy (SO) data from planned surveys, and presence-background (PB) data from opportunistic surveys and other sources. SO surveys give high quality data about presences and absences of the species in a particular area. However, due to their high cost, they often cover a smaller area relative to PB data, and are usually not representative of the geographic range of a species. In contrast, PB data is plentiful, covers a larger area, but is less reliable due to the lack of information on species absences, and is usually characterised by biased sampling. Here we present a new approach for species distribution modelling that integrates these two data types. We have used an inhomogeneous Poisson point process as the basis for constructing an integrated SDM that fits both PB and SO data simultaneously. It is the first implementation of an Integrated SO–PB Model which uses repeated survey occupancy data and also incorporates detection probability. The Integrated Model's performance was evaluated, using simulated data and compared to approaches using PB or SO data alone. It was found to be superior, improving the predictions of species spatial distributions, even when SO data is sparse and collected in a limited area. The Integrated Model was also found effective when environmental covariates were significantly correlated. Our method was demonstrated with real SO and PB data for the Yellow-bellied glider ( Petaurus australis ) in south-eastern Australia, with the predictive performance of the Integrated Model again found to be superior. PB models are known to produce biased estimates of species occupancy or abundance. The small sample size of SO datasets often results in poor out-of-sample predictions. Integrated models combine data from these two sources, providing superior predictions of species abundance compared to using either data source alone. Unlike conventional SDMs which have restrictive scale-dependence in their predictions, our Integrated Model is based on a point process model and has no such scale-dependency. It may be used for predictions of abundance at any spatial-scale while still maintaining the underlying relationship between abundance and area.

Methods in Ecology and Evolution↗