Dissolved Oxygen Prediction Model for the Yangtze River Estuary Basin Using IPSO-LSSVM

Yong-guo, LI; Li, X; Xu, Caiyin; Tang, Xuan

doi:10.3390/w15122206

Cited by 8 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Observed or ground-truth data from conventional field sampling should be used to complement RS measurements: this is important to ensure proper calibration and validation of RS models. Statistical or evaluation metrics such as the R 2 , Percent Bias (PBIAS), MAE, NSE, RMSE, and the ratio of the RMSE to standard deviation which have been widely used should be utilized to evaluate developed models to improve their accuracy and robustness [2,128,145,[271][272][273][274][275][276][277]. 4.…”

Section: Conclusion and Recommendationsmentioning

confidence: 99%

Measurement of Total Dissolved Solids and Total Suspended Solids in Water Systems: A Review of the Issues, Conventional, and Remote Sensing Techniques

et al. 2023

View full text Add to dashboard Cite

This study provides a comprehensive review of the efforts utilized in the measurement of water quality parameters (WQPs) with a focus on total dissolved solids (TDS) and total suspended solids (TSS). The current method used in the measurement of TDS and TSS includes conventional field and gravimetric approaches. These methods are limited due to the associated cost and labor, and limited spatial coverages. Remote Sensing (RS) applications have, however, been used over the past few decades as an alternative to overcome these limitations. Although they also present underlying atmospheric interferences in images, radiometric and spectral resolution issues. Studies of these WQPs with RS, therefore, require the knowledge and utilization of the best mechanisms. The use of RS for retrieval of TDS, TSS, and their forms has been explored in many studies using images from airborne sensors onboard unmanned aerial vehicles (UAVs) and satellite sensors such as those onboard the Landsat, Sentinel-2, Aqua, and Terra platforms. The images and their spectral properties serve as inputs for deep learning analysis and statistical, and machine learning models. Methods used to retrieve these WQP measurements are dependent on the optical properties of the inland water bodies. While TSS is an optically active parameter, TDS is optically inactive with a low signal–noise ratio. The detection of TDS in the visible, near-infrared, and infrared bands is due to some process that (usually) co-occurs with changes in the TDS that is affecting a WQP that is optically active. This study revealed significant improvements in incorporating RS and conventional approaches in estimating WQPs. The findings reveal that improved spatiotemporal resolution has the potential to effectively detect changes in the WQPs. For effective monitoring of TDS and TSS using RS, we recommend employing atmospheric correction mechanisms to reduce image atmospheric interference, exploration of the fusion of optical and microwave bands, high-resolution hyperspectral images, utilization of ML and deep learning models, calibration and validation using observed data measured from conventional methods. Further studies could focus on the development of new technology and sensors using UAVs and satellite images to produce real-time in situ monitoring of TDS and TSS. The findings presented in this review aid in consolidating understanding and advancement of TDS and TSS measurements in a single repository thereby offering stakeholders, researchers, decision-makers, and regulatory bodies a go-to information resource to enhance their monitoring efforts and mitigation of water quality impairments.

show abstract