Codes and data file for "Declining of atmospheric wet deposition of dissolved organic carbon in China after 2017"

It documents model performance, including cross-validated metrics, and identifies the final model (Model 41: rainfall + SO₂ + CO + O₃) used for subsequent national-scale prediction.</p><h3 dir="ltr">2. <code>comparable model results plus data for Figure 4</code></h3><p dir="ltr"><b>Purpose:</b> Results from comparable models and the input/output data used to generate Figure 4.</p><p dir="ltr">This file contains the performance and prediction results of the models that showed comparable predictive performance to the final model, together with the data required for the national-scale spatial prediction shown in Figure 4.

These results are used to evaluate the robustness of model selection and to generate the spatial distribution of wet DOC deposition across China.</p><h3 dir="ltr">3. <code>national scale prediction and monte carlo simulation</code></h3><p dir="ltr"><b>Purpose:</b> National-scale DOC deposition estimation and uncertainty analysis.</p><p dir="ltr">This file contains the station-level predictions, annual aggregation, spatial interpolation, and Monte Carlo simulation results used to estimate annual wet DOC deposition across China from 2015 to 2023.

The Monte Carlo analysis propagates model and spatial-interpolation uncertainty to provide uncertainty ranges for the national annual DOC deposition estimates.</p><h3 dir="ltr">4. <code>LOSO and LORO analysis</code></h3><p dir="ltr"><b>Purpose:</b> Evaluation of model extrapolation robustness.</p><p dir="ltr">This file contains the results of leave-one-site-out (LOSO) and leave-one-region-out (LORO) validation analyses.

These analyses evaluate the robustness of the final model when predicting at spatial locations or regions that are excluded from model fitting, providing an assessment of its suitability for national-scale extrapolation.</p><h3 dir="ltr">5. <code>year-to-year heatmap</code></h3><p dir="ltr"><b>Purpose:</b> Analysis of interannual changes in wet DOC deposition.</p><p dir="ltr">This file contains the year-to-year changes in modeled DOC deposition and the corresponding changes in rainfall and atmospheric air-quality variables.

It is used to generate the year-to-year correlation heatmap and to evaluate how precipitation and air-quality conditions are associated with spatial differences in interannual variability.</p><h3 dir="ltr">6. <code>fix_air quality_counterfactual</code></h3><p dir="ltr"><b>Purpose:</b> Counterfactual analysis with air-quality variables held constant.</p><p dir="ltr">This file contains the results of the counterfactual simulation in which air-quality parameters are held constant while rainfall is allowed to vary among years.

The analysis is used to assess how well changes in rainfall alone reproduce the reconstructed temporal trajectory of national wet DOC deposition.</p><h3 dir="ltr">7. <code>fix rainfall_counterfactual</code></h3><p dir="ltr"><b>Purpose:</b> Counterfactual analysis with rainfall held constant.</p><p dir="ltr">This file contains the results of the counterfactual simulation in which rainfall is held constant while air-quality parameters are allowed to vary among years.

The analysis is used to assess how closely changes in atmospheric air-quality conditions reproduce the temporal trajectory of national wet DOC deposition.</p><h3 dir="ltr">8. <code>rainwater DOC data (training dataset)</code></h3><p dir="ltr"><b>Purpose:</b> Original rainwater DOC observations used for model development.</p><p dir="ltr">This file contains the rainwater DOC concentration/deposition observations and associated environmental variables used to develop and calibrate the GAMs.

The dataset represents the observational training dataset used to establish the relationship between wet DOC deposition and rainfall and atmospheric air-quality conditions.</p><h3 dir="ltr">9. <code>Supplementary_national air quality</code></h3><p dir="ltr"><b>Purpose:</b> National-scale environmental input data and model predictions.</p><p dir="ltr">This file contains the monthly rainfall and atmospheric air-quality data (PM₂.₅, PM₁₀, SO₂, NO₂, CO, and O₃) for 323 sites/cities across China during 2015–2023, together with the corresponding model prediction results.

These data provide the environmental inputs for national-scale application of the final model and support the generation of the spatial and temporal analyses presented in the main text and Supplementary Information.</p>