Research ArticleENVIRONMENTAL STUDIES

China’s improving inland surface water quality since 2003

See allHide authors and affiliations

Science Advances  03 Jan 2020:
Vol. 6, no. 1, eaau3798
DOI: 10.1126/sciadv.aau3798
  • Fig. 1 Geographical distributions of major river basins and site-level quality of inland surface waters in China.

    Different colors overlaid on the sampling sites (dots, n = 2630) indicate regionally varying quality levels of annual mean concentrations of water quality parameters in 2017. (A) COD. (B) NH4+-N. (C) DO. Levels I, II, and III generally signify protected potable water sources (50). Detailed information regarding 10 target basins is summarized in table S1. (D) Probability density of site-level monthly mean concentrations of three water quality indicators in 2003 (green), 2010 (blue), and 2017 (red). Dashed lines represent the arithmetic mean values.

  • Fig. 2 Country-level trends in annual estimates of water quality measurements and quality levels of inland surface water in China from 2003 to 2017.

    (A to C) Country-level trends in five time series of COD, NH4+-N, and DO concentrations. (D to F) Interannual variability in percentages of five water quality levels (as defined in Fig. 1) limited by site-level records of COD, NH4+-N, and DO concentrations, respectively. Solid lines represent estimated linear trends in concentrations in (A) to (C) and the summed percentage of water quality levels I + II + III in (D) to (F), respectively. *** stands for statistical significance at the 0.01 level. Detailed statistical results for (A) to (F) are summarized in tables S2 and S3.

  • Fig. 3 Country-level dynamics in anthropogenic pollution discharges from different sectors and their relative impacts on interannual variances in observed COD and NH4+-N concentrations of China’s inland water during 2003–2017.

    (A) Pollution discharges measured by COD from different sectors. (B) NH4+-N discharges from different sectors. (C and D) Relative impacts (here estimated by the loadings of the first principal component; see Materials and Methods) of different pollution sources to interannual variances in observed COD and NH4+-N concentrations, respectively. Solid lines in (A) and (B) are fitted local trends using locally weighted scatterplot smoothing with 0.8 bandwidth. Gray areas represent the 95% confidence intervals estimated by the bootstrap sampling method. Both (C) and (D) are the analyzed results for country-level Qm and Qa time series (Fig. 2, A and B). Positive values in (C) and (D) indicate a positive effect of reduced sector-specific pollution discharges on the declination in observed concentrations of either COD or NH4+-N and vice-versa.

  • Fig. 4 Spatial distributions of anthropogenic pollution discharges (in Gg year−1 per 100 km2) and temporal changes between 2003 and 2015.

    (A) Pollution discharge measured by COD in 2015. (B) Changes in pollution discharge measured by COD between 2003 and 2015. (C) NH4+-N discharge in 2015. (D) Changes in NH4+-N discharge between 2003 and 2015. Spatial estimates are based on assembled anthropogenic pollution discharge datasets involved in this study and satellite-derived land-use maps and performed on 10 km–by–10 km grids (see Materials and Methods).

  • Fig. 5 Quantitative relationships between annual estimates of observed regional-level COD and NH4+-N concentrations and anthropogenic pollution discharge loadings (with respect to the volume of local surface water resources) among 10 major river basins circa 2015.

    (A) COD. (B) NH4+-N. Both were fitted by a power law model. Red and blue dots and lines represent observed data and fitted results for Qm and Qa (here, both are annual average concentrations in 2015–2017), respectively. Dashed lines link the paired observations of Qm and Qa in the same basin. Basins are labeled according to Qm. Gray areas represent the 95% confidence intervals (CI).

  • Fig. 6 Similarity of impact factors for 10 major river basins circa 2015.

    (A) Cluster dendrogram of impact factors of water quality based on the distance metrics over 10 basins. (B) Hierarchical clustering of 10 basins based on the distance metrics among multiple impact factors. WAT, surface water resource per unit area; PRE, average annual precipitation; GRE, the percent of green cover (including forest land and grassland with >20% of cover degree); POP, population density per unit of surface water resource; GDP, gross domestic product density per unit of surface water resource; INV, the proportion of environmental investment accounting for GDP. All data for clustering analysis are listed in table S5.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/1/eaau3798/DC1

    Supplementary Text

    Fig. S1. National-level changes in multiple driving factors of inland water quality in China during the period 2003–2015.

    Fig. S2. Observed changes in water quality measures for five time series across 10 basins during the period 2003–2017.

    Fig. S3. Interannual variability in percentages of five water quality levels limited by site-level records of water quality measures for 10 basins during the period 2003–2017.

    Fig. S4. Changes in anthropogenic pollution discharges from different sectors across 10 basins during the period 2003–2017.

    Fig. S5. Changes in the volume of environmental investment and the rate of access to sanitary toilets for rural residents across 10 basins during the period 2003–2015.

    Fig. S6. The relationships between environmental investments and declining pollution emissions and changes in the discharge levels between 2003 and 2015 across 10 basins.

    Table S1. Summary of the distinguishing characteristics of 10 basins delineated in this study.

    Table S2. Summary of analysis results of trends in water quality indicators during the period 2003–2017 across China and 10 basins.

    Table S3. Summary of analysis results of trends in the proportion of water quality levels I + II + III for water quality indicators for the period 2003–2017 across China and 10 basins.

    Table S4. Coefficients of the first component derived from PCA of loading sources and used for the explained variances estimated by type II ANOVA in both the Qm and Qa time series of COD and NH4+-N concentrations from 2003–2017 across 10 basins.

    Table S5. Summary of impact factors of local inland surface water quality across 10 basins.

  • Supplementary Materials

    This PDF file includes:

    • Supplementary Text
    • Fig. S1. National-level changes in multiple driving factors of inland water quality in China during the period 2003–2015.
    • Fig. S2. Observed changes in water quality measures for five time series across 10 basins during the period 2003–2017.
    • Fig. S3. Interannual variability in percentages of five water quality levels limited by site-level records of water quality measures for 10 basins during the period 2003–2017.
    • Fig. S4. Changes in anthropogenic pollution discharges from different sectors across 10 basins during the period 2003–2017.
    • Fig. S5. Changes in the volume of environmental investment and the rate of access to sanitary toilets for rural residents across 10 basins during the period 2003–2015.
    • Fig. S6. The relationships between environmental investments and declining pollution emissions and changes in the discharge levels between 2003 and 2015 across 10 basins.
    • Table S1. Summary of the distinguishing characteristics of 10 basins delineated in this study.
    • Table S2. Summary of analysis results of trends in water quality indicators during the period 2003–2017 across China and 10 basins.
    • Table S3. Summary of analysis results of trends in the proportion of water quality levels I + II + III for water quality indicators for the period 2003–2017 across China and 10 basins.
    • Table S4. Coefficients of the first component derived from PCA of loading sources and used for the explained variances estimated by type II ANOVA in both the Qm and Qa time series of COD and NH4+-N concentrations from 2003–2017 across 10 basins.
    • Table S5. Summary of impact factors of local inland surface water quality across 10 basins.

    Download PDF

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article