Atmospheric wet deposition has attracted much attention because of its removal function of aerosol particles and impact on ecosystem. It is a complex multi-component process due to various pollutants emission sources and reaction processes. However, traditional studies have focused on single components, ignoring the internal relationships and the comprehensive wet deposition level; this has hindered the comprehensive understanding and the assessment of ecological effects of deposition. Here, based on the monitoring of wet deposition across 52 stations during 2013–2018 in China, we used two novel multidimensional methods to investigate multi-component atmospheric wet deposition: network analysis and comprehensive index. Network analysis is a new method that can help us better explore the overall relationship between multiple components. The tighter deposition networks, the stronger relationship between different components, meaning that their source are more similar. We find that with the improvement of human development level and the diversification of energy structure, the deposition relationship networks will become looser. Comprehensive index is a method to realize the fusion of multiple parameters, which can help us explore the comprehensive effect of multi-component atmospheric deposition on ecosystem. We calculated the deposition comprehensive index for different ecological regions and found that it was strongly influenced by human activities (energy consumption, fertilization, and vehicle ownership) and precipitation. The findings presented here provide a new perspective for understanding multi-component atmospheric deposition and potential suggestions for evaluating its comprehensive ecological effects. 

Fifty-two ecological stations were selected to monitor the multi-component atmospheric wet deposition from 2013 to 2018 from the ChinaWD . These stations are distributed across 22 provinces. They were divided into eight ecological regions according to climate and vegetation type: northeast, Inner Mongolia, central China, north China, south China, northwest, southwest, and Qinghai-Tibet.

Following the long-term monitoring guide from the ChinaWD, the samples were collected manually using plastic buckets installed 1.5 m above the ground when precipitation occurred. After each precipitation event, samples were transferred to polyethylene plastic bottles (which were cleaned with deionized water three times and air-dried prior to sampling) and immediately stored at –20 °C to prevent any possible transformation by microbes. At the end of each month, these subsamples were mixed to prepare a representative monthly sample.

In the laboratory, suspended particles were removed from samples through a 0.45 μm microporous membrane; Then, the concentrations of NH₄⁺ and NO₃^–were determined using a continuous flow analyzer (FUTURA, Alliance Instruments, France); Next, an inductively coupled plasma optical emission spectrometer (ICP-OES; Optima 5300DV, PerkinElmer, America) was used to measure the concentrations of base cations (Ca²⁺, K⁺, Na⁺, and Mg²⁺), trace metal ions (Zn, Ni, Cu, Mn, Fe, and Pb), and SO₄^2–. Meanwhile, Cl^–, F^–, and NO₂^– concentrations were measured using an ion chromagraph (ICS1100), with data for these ions available since 2018.

The fluxes of atmospheric wet deposition were calculated as following:

where D_j (kmol·ha^–1·yr^–1) is the annual deposition flux of the j^th deposition component, C_ij (mg·L^–1) is the monthly concentration of the j^th component, P_i (mm) is the monthly precipitation, it is measured in every sites, n is the number of months, and M_j is the molar mass of the j^th component.

The wet depositions of different regions were obtained by averaging the deposition fluxes of the observation points in each region. The wet deposition fluxes for the 22 provinces were obtained in the same way. In addition, to account for the heterogeneity across different regions the mean deposition in China was calculated using an area-weighted method.

Complex network analysis is an effective method for evaluating the relationships between different dimensions;. Multidimensional networks are composed of nodes and edges. To create a deposition relationship network (DRN), here all the deposition components were delineated as nodes, and the relationships between the different components were defined as edges. First, the correlation coefficient matrices of the different deposition components were calculated. Then, a threshold of marked pairwise correlations was used, with p < 0.05 showing a significant difference. Other relationships were set to zero, yielding the adjacency matrix A= [ai,j] with ai,j ϵ [0,1]. Finally, the improved package “igraph” in R was employed to visualize the DRN.After creating each DRN, a suitable parameter was selected to quantify the overall topology characteristics.  The edge density (ED) describes the tightness of the connected edges between nodes in a network, that is, the proportion of actual connections among components out of all possible connections; it can be calculated using Eq. (2) :

where L is the number of actual edges in a network, and n is the number of node components. In general, it is considered that some components exhibit high correlations owing to their similar origins and interactions. Therefore, a DRN with a high ED implies that the sources of deposition in the region in question are relatively single, while a low ED reflects that the sources are more diversified.

Deposition comprehensive index (DCI) was assessed according to Fanin et al. (2018) as defined by Eq. (3):

where DCI_j  is the DCI in region j, x_ij  is the wet deposition of the i^th component in region j; min⁡(x_j)  and maxx_j  are the maximum and minimum of the i^th component in all areas, respectively; and n is the number of components. In brief, the different deposition component data were normalized to a scale of 0 to 1, and the average of these standardized values was obtained as the DCI for each region. A high DCI indicates that multi-component atmospheric deposition has greater potential impact on ecosystems in this region than in other regions, while a low DCI indicates less potential impact .