Industrial Wastewater Discharge Retrieval Based on Stable Nighttime Light Imagery in China from 1992 to 2010

Zhaofu Li, Hongyu Liu, Chuan Luo, Panpan Li, Hengpeng Li, Zhengqin XiongRemote 2014 · DOI

TL;DR

This paper pioneers the use of DMSP/OLS nighttime light imagery to estimate industrial wastewater (IW) discharge across China from 1992 to 2010. By establishing strong correlations between nighttime light intensity and IW discharge, the study provides a valuable, high-resolution dataset for understanding pollution trends and informing water quality models.

What Problem Does This Paper Address?

Industrial wastewater (IW) is a significant source of water pollution, posing a growing threat to aquatic ecosystems and human health. Accurately assessing the spatial and temporal distribution of IW discharge is crucial for effective water resource management and environmental modeling. However, traditional methods rely on administrative-scale statistical data, which are often coarse, inconsistent, and not directly applicable to hydrological processes that occur at the basin or watershed level. This disconnect limits our ability to understand and predict water quality dynamics. The paper aims to bridge this gap by leveraging remote sensing data to provide a more granular and consistent estimation of IW discharge.

Methods

The core of this study lies in establishing a quantitative relationship between nighttime light (NTL) data and industrial wastewater discharge. The researchers utilized stable nighttime light imagery from the Defense Meteorological Satellite Program's Operational Linescan System (DMSP/OLS) for the period 1992-2010.

  1. Data Preparation: DMSP/OLS NTL data, processed into 30-arc-second grids (approximately 0.55 km resolution), were aggregated to a 1 km x 1 km resolution. Provincial boundaries were overlaid, and the average digital number (DN) value per unit area (DNP) was calculated for each province. Industrial wastewater discharge per unit area (IWP) was derived by dividing the total industrial wastewater (TIW) by the province's area.
  2. Model Development: Linear regression models were employed to establish the relationship between DNP and IWP for each province and year. Various models (exponential, power, quadratic) were tested, with the linear model ($y = kx + b$) showing the best performance.
  3. Validation: The predicted total industrial wastewater (PTIW) derived from the regression models was compared with official statistical total industrial wastewater (STIW) data to assess accuracy. The relative difference (RD) was calculated as a metric for error.

Datasets

  • DMSP/OLS Nighttime Light (NTL) Data: Version 4 stable light data from 1992 to 2010, obtained from the National Geophysical Data Center (NGDC) at NOAA. This dataset provides a measure of artificial light emitted from the Earth's surface.
  • China Statistical Yearbooks (1993-2011): These yearbooks provided the official statistical data for total industrial wastewater (TIW) and gross industrial production (GIP) for each province from 1992 to 2010, compiled by the National Bureau of Statistics of China.

Key Insights and Ideas

  • NTL as a Proxy for IW Discharge: The study demonstrates that nighttime light intensity, a proxy for human activity and economic development, is a surprisingly strong indicator of industrial wastewater discharge. This opens up new avenues for monitoring pollution in data-scarce regions.
  • Spatial Granularity: By using gridded NTL data, the study moves beyond administrative boundaries to provide a more spatially explicit understanding of IW discharge, which is crucial for localized environmental management.
  • Temporal Trends Reveal Shifting Patterns: The analysis reveals a significant trend of expanding IW discharge coverage alongside a decreasing intensity of discharge per unit area over two decades. This suggests a shift towards more diffuse pollution or improved treatment efficiency in heavily industrialized areas.
  • Consistency with Economic Development: The NTL-based estimates generally align with economic development patterns, with higher discharge concentrated in economically developed eastern and coastal regions.

Key Results

  • Strong Correlations: Significant positive correlations were found between DNP and IWP across provinces and years. Pearson correlation coefficients (r) for DNP vs. IWP were consistently above 0.85. Similar strong correlations were observed for chemical oxygen demand (IWCODP) and ammonia nitrogen (IWNH3NP) discharge.
  • Reliable Estimation: The fitted linear regression models provided a dependable estimation of IW discharge with a total error of -8.2% in 2010 when compared to statistical data.
  • Spatial Distribution: IW discharge is spatially concentrated in eastern, northeastern, and southeastern China, particularly in major urban and coastal areas like the Jing-Jin-Ji circle, Yangtze River Delta, and Pearl River Delta.
  • Temporal Evolution: From 1992 to 2010, the area affected by IW discharge expanded significantly, while the intensity (discharge per unit area) decreased. For instance, the maximum IWP decreased from 50.071 × 10⁴ t/km² in 1992 to 6.332 × 10⁴ t/km² in 2010.

Limitations and Future Work

  • Model Accuracy Variations: While generally reliable, the model exhibited significant errors in specific provinces. The authors attribute these discrepancies to regional differences in economic development and industrial productivity efficiency. Provinces with a higher percentage of IW discharge in their total national discharge showed more accurate estimates.
  • NTL Data Limitations: The DMSP/OLS data has limitations, including saturation in highly lit areas and potential misinterpretation of light sources (e.g., gas flares, fires). The study acknowledges that the relationship between NTL and IW discharge efficiency is complex.
  • Future Work: The authors suggest improving prediction accuracy by incorporating other factors that influence IW discharge into the regression models. They also highlight the need to integrate industrial wastewater data with domestic sewage and agricultural non-point source pollution for a more comprehensive water environment pollution assessment. Future research could also explore the use of newer, higher-resolution NTL datasets.

Summary updated

Abstract

Industrial wastewater (IW) discharge, which is a known point source of pollution, is a major water pollution source. Increasing IW discharge has imposed considerable pressure on regional or global water environments. It is important to estimate the IW distribution in grid units to improve basin-scale hydrological processes and water quality modeling. For the first time, we use the nighttime light imagery produced by the Defense Meteorological Satellite Program’s Operational Linescan System (DMSP/OLS) to estimate the spatial and temporal variations in the IW distribution from 1992 to 2010 in China. The digital number values per unit area (DNP) of each stable light image were calculated using nighttime light imagery and were regressed against the IW per unit area (IWP) to estimate the total industrial wastewater (TIW) for each province. The results indicated strong positive correlations between the DNP and the IWP for each province during different years. The fitted linear regression models were used to estimate IW discharge in China with reliable accuracy. The IW estimation using the satellite data was consistent with the statistical results. The results also revealed that the IW discharge coverage expanded, whereas the IW discharge intensity decreased from 1992 to 2010 in China.
BibTeX
@inproceedings{li2014industrial,
  author = {Zhaofu Li and Hongyu Liu and Chuan Luo and Panpan Li and Hengpeng Li and Zhengqin Xiong},
  title = {{Industrial Wastewater Discharge Retrieval Based on Stable Nighttime Light Imagery in China from 1992 to 2010}},
  booktitle = {Remote},
  year = {2014},
  doi = {10.3390/rs6087566},
}
Figure