A global dataset of atmospheric ⁷Be and ²¹⁰Pb measurements: annual air concentration and depositional flux

⁷Be and ²¹⁰Pb air concentration and depositional flux data provide key information on the origins and movements of air masses, as well as atmospheric deposition processes and residence time of aerosols. After their deposition onto the Earth's surface, they are utilized for tracing soil redistribution processes on land, particle dynamics in aquatic systems, and mixing processes in open ocean. Here we present a global dataset of air concentration and depositional flux measurements of atmospheric ⁷Be and ²¹⁰Pb made by a large number of global research communities. Data were collected from published papers between 1955 and early 2020. It includes the annual surface air concentration data of ⁷Be from 367 sites and ²¹⁰Pb from 270 sites, the annual depositional flux data of ⁷Be from 279 sites and ²¹⁰Pb from 602 sites. When available, appropriate metadata have also been summarized, including geographic location, sampling date, methodology, annual precipitation, and references. The dataset is archived at https://doi.org/10.5281/zenodo.4785136 (Zhang et al., 2021) and is freely available for the scientific community. The purpose of this paper is to provide an overview of the scope and nature of this dataset and its potential utility as baseline data for future research.

1 Introduction

Naturally occurring beryllium-7 (⁷Be, $T_{\mathrm{1}/\mathrm{2}}$: 53.3 d) and lead-210 (²¹⁰Pb, $T_{\mathrm{1}/\mathrm{2}}$: 22.3 years) have been widely utilized as tracers to investigate Earth's surface and atmospheric processes (Huh et al., 2006; Du et al., 2012). ⁷Be, a cosmogenic radionuclide, is produced by the spallation of oxygen and nitrogen nuclei by cosmic rays in the stratosphere and upper troposphere (Lal et al., 1958). The production rate of ⁷Be has negligible dependence on longitude or season but depends on the altitude, latitude, and the ∼ 11-year solar cycle (Koch et al., 1996; Liu et al., 2001; Su et al., 2003). A major fraction of ⁷Be (67 %) production takes place in the stratosphere, but it does not readily reach the troposphere, except during spring when seasonal thinning of the tropopause folds near the jet stream occurs at midlatitudes (Lal and Peters, 1967; Danielsen, 1968). Thus, ⁷Be flux to the Earth's surface varies with latitude and season (Lal and Peters, 1967; Koch and Mann, 1996). ²¹⁰Pb, a progeny of ²²²Rn in the ²³⁸U series, is derived mostly (> 99 %) from the radioactive decay of ²²²Rn. Most of the atmospheric ²¹⁰Pb is derived from atmospheric radon. The global ²²²Rn flux from continents ranged from 1300 to 1800 Bq m⁻² d⁻¹, while 2–21 Bq m⁻² d⁻¹ was reported for oceanic areas (Wilkening and Clements, 1975; Nazaroff, 1992). Vertical profiles of ²²²Rn in the atmosphere indicate the highest concentrations occur in the continental boundary layer (CBL, 3–8 Bq m⁻³), while activity that is lower by an order of magnitude (∼ 40 mBq m⁻³) occurs near the tropopause, with decreasing activity with increasing altitude from the CBL (Moore et al., 1977; Liu et al., 1984; Kritz et al., 1993). Consequently, the atmospheric concentration of ²¹⁰Pb decreases with increasing altitude and is strongly controlled by the land–sea distribution pattern. After formation, ⁷Be, ²¹⁰Pb, and short-lived ²²²Rn progeny are all rapidly and irreversibly attached to aerosol particles (Winkler et al., 1998; Elsässer et al., 2011). Subsequently, the fate of ⁷Be and ²¹⁰Pb is closely linked to that of aerosols. Most of ⁷Be and ²¹⁰Pb are in accumulation-mode aerosol particles with an aerodynamic diameter of a few hundred nanometers (Ioannidou and Paatero, 2014; Paatero et al., 2017). Therefore, they are deposited onto the Earth's surface primarily by precipitation because accumulation-mode aerosol particles are too small for gravitational settling and removal and too large to be deposited by Brownian motion.

Owing to their distinctly different source terms but well-known source distributions and similar tropospheric physicochemical behavior, ⁷Be and ²¹⁰Pb have been widely utilized as powerful atmospheric tracers for studying the origin of air masses (e.g., Graustein and Turekian, 1996; Zheng et al., 2005; Likuku et al., 2006; Dueñas et al., 2011; Lozano et al., 2012), vertical exchange and horizontal transport processes (e.g., Arimoto et al., 1999; Lee et al., 2007; Rastogi and Sarin, 2008; Tositti et al., 2014), deposition velocities and washout ratios of aerosols (e.g., Todd et al., 1989; McNeary and Baskaran, 2003; Dueñas et al., 2005; Lozano et al., 2011; Mohan et al., 2019), and the behavior and fate of analog species (e.g., Crecelius, 1981; Mattsson, 1988; Prospero et al., 1995; Lamborg et al., 2013). Following their deposition on the Earth's surface, both ⁷Be and ²¹⁰Pb are strongly attached to soils, which make them useful for assessing soil erosion rates from episodic to multi-decadal timescales (e.g., Wallbrink and Murry, 1993; Walling and He, 1999; Blake et al., 1999; Walling et al., 1999; Wilson et al., 2003; Mabit et al., 2008, 2014). In aquatic environments, ²¹⁰Pb is most widely used for dating recent sediments (e.g., Appleby, 2008). Meanwhile, ⁷Be and ²¹⁰Pb are also widely used as tracers of sediment source identification and particle dynamics in rivers (e.g., Bonniwell et al., 1999; Matisoff et al., 2005; Jweda et al., 2008; Mudbidre et al., 2014; Baskaran et al., 2020), lakes (e.g., Dominik et al., 1987; Schuler et al., 1991; Vogler et al., 1996), and estuaries and coasts (e.g., Baskaran et al., 1997; Huang et al., 2013; Wang et al., 2016). ⁷Be deposited on open ocean is further used as a tracer for diagnosing ocean ventilation and subduction (Kadko, 2000; Kadko and Olson, 1996), inferring upwelling rates (Kadko and Johns, 2011), and estimating the deposition of trace metals (Kadko et al., 2015; Shelley et al., 2017; Buck et al., 2019).

There have been numerous published datasets with concentrations and depositional fluxes of ²¹⁰Pb and ⁷Be (directly and indirectly) over the past few decades, particularly under national or international monitoring programs, such as the Environmental Measurements Laboratory (EML) Surface Air Sampling Program (Feely et al., 1989), the Sea-Air Exchange (SEAREX) program (Uematsu et al., 1994), the Finnish Meteorological Institute monitoring program (Paatero et al., 2015), the Radioactivity Environmental Monitoring (REM) network (Hernandez-Ceballos et al., 2015), and the International Monitoring System (IMS) operated by the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) (Terzi and Kalinowski, 2017). It is valuable to compile all of these existing data, including those measured in case studies, along with appropriate metadata, in one place for facilitating further data analysis and geological, geochemical, and geophysical applications.

Although several datasets of air concentrations and depositional fluxes of ⁷Be (Bleichrodt, 1978; Brost et al., 1991) and ²¹⁰Pb (Rangarajan et al., 1986; Preiss et al., 1996) have been published, unfortunately, many of the published data are not readily available in a hard data format (e.g., data table), and retrieval of data from published figures is often not precise and can be challenging. To date, only one dataset was published that compiled ⁷Be and ²¹⁰Pb together (Persson, 2015), but it contained limited data. Therefore, the focus of the present work is to build a new and comprehensive dataset. This dataset is the result of many scientists' efforts in generating the data.

2 Methods

2.1 ⁷Be and ²¹⁰Pb air concentrations measurement methodology

Aerosol samples are usually collected by filtering a high volume of air, typically 1.4–1.5 m³ min⁻¹, collected typically for a day, through paper filters. Commonly used aerosol collecting equipment include the following instruments: ASS 500 (CLRP Warsaw, Germany), Anderson PM10 (Anderson Ltd., USA), Snow White (Senya Ltd., Finland), and HV-1000F (Shibata Co. Ltd., Japan). The preferred filter membrane materials include glass fiber, cellulose nitrate or acetate, polypropylene fiber, and quartz fiber. The collection efficiency, the percentage of the particles in the air stream that are collected by the air filter, depends on the aerodynamic diameter of aerosol particles and the filter face velocity (average flow velocity of air into the filter) of the airflow. Although the collection efficiency depends on the distribution of particle sizes, generally the collection efficiency varies between 80 % and 100 % for different filter materials. This was shown by overlapping two filters in tandem and comparing the concentrations separately in the top and bottom filters. The sampling frequency is usually set from daily to monthly, with a typical collection time of ∼ 24 h, corresponding to ∼ 2000 m³ air. The ⁷Be and ²¹⁰Pb trapped on filters can both be analyzed simultaneously by gamma spectrometry (e.g., McNeary and Baskaran, 2003; Bourcier et al., 2011; Lozano et al., 2012; Mohan et al., 2018), while ²¹⁰Pb can also be analyzed by beta counting of its daughter ²¹⁰Bi (e.g., Joshi et al., 1969; Poet et al., 1972; Daish et al., 2005) or via alpha counting of in-grown ²¹⁰Po from the decay of ²¹⁰Pb (e.g., Turekian and Cochran, 1981; Mattsson et al., 1996; Marx et al., 2005).

2.2 ⁷Be and ²¹⁰Pb depositional fluxes measurement methodology

The atmospheric depositional fluxes of ⁷Be and ²¹⁰Pb are commonly measured directly by using rain collectors, such as polyethylene drums or buckets and stainless steel containers, and indirectly by natural archives (e.g., soils, lichens, mosses, snow and ice cores, salt marsh sediments). The direct collecting method is the most reliable technique for the measurement of annual ⁷Be and ²¹⁰Pb depositional fluxes, and this technique is useful in collecting short timescale (daily, weekly, and monthly) depositional fluxes. In contrast, using natural archives avoids the labor- and time-intensive measurement of the ⁷Be and ²¹⁰Pb concentration in precipitation and can serve as a complement to fill regional gaps, especially in remote areas. However, these archives are susceptible to being affected by natural processes and anthropogenic activities; thus, the sampling location of these archives should be restricted to undisturbed areas.

2.2.1 Direct ⁷Be and ²¹⁰Pb flux measurements

Rain collectors are usually placed on the roof of a building so as to prevent contamination from resuspended dust from the ground. Care should be taken to ensure direct overhead atmospheric deposition is collected and that there is no shadowing effect from adjacent structures or buildings or a funneling effect in sample collection. Atmospheric aerosols can be removed not only by precipitation-scavenging but also by settling under the influence of gravitational or electrostatic forces. In most cases, the collectors are continuously exposed over a long enough period, and the bulk (wet + dry) depositional samples are collected periodically (e.g., Baskaran et al., 1993; Hirose et al., 2004; Baskaran and Swarzenski, 2007; Lozano et al., 2011; Du et al., 2015). Fluxes obtained by this method yield the best estimate of the depositional flux. Sometimes, fallout ⁷Be and ²¹⁰Pb samples are collected only during rainfall, and concentration is measured in the individual rainwater sample (e.g., Cho et al., 2011; Chae and Kim, 2019; Du et al., 2020). In this case, only the bulk depositional flux is obtained for the duration of collection. In rare cases, when only a mean concentration of ²¹⁰Pb and/or ⁷Be in rainwater is available, the wet flux can be estimated by multiplying by the annual precipitation (Peirson et al., 1966).

Most of the time, the volume of rainwater sample is large and cannot be directly counted in a gamma-ray spectrometer for the simultaneous measurements of ⁷Be and ²¹⁰Pb, in which case a preconcentration of the sample is required. Since both ⁷Be and ²¹⁰Pb have a strong affinity for solid surfaces, it is strongly recommended to add stable Be (commonly 1–5 mg) and stable Pb (typically 5–20 mg) in about 1 L of 1 M HCl to the rain collector prior to deployment. Alternatively, the spikes can also be immediately added after sample collection, followed by rinsing of rain collector with 1 L of 1 M HCl rinsing twice and combining the rinses with the collected rainwater. An earlier critical review of earlier atmospheric depositional flux studies by Lal et al. (1979) showed that loss of ⁷Be and ²¹⁰Pb by sorption onto rain collector walls was observed when pre-acidification of the collector was not done, resulting in the underestimation of depositional flux. In the case of preconcentration by ferric chloride precipitation method, due to variable scavenging efficiency of ⁷Be and ²¹⁰Pb during the preconcentration method, it is required to add stable Pb and Be as a yield tracer (e.g., Baskaran et al., 1993). The best chemical procedure to obtain high-quality data is to add acid and stable Be and Pb careers with 1 L of 1 M HCl to the rain collector prior to the start of the sample collection. The final calculation for depositional fluxes would involve chemical yield for ⁷Be and ²¹⁰Pb and appropriate decay corrections, as outlined in Baskaran et al. (1993) and Du et al. (2015).

2.2.2 Indirect ⁷Be flux measurements

Measurements of ⁷Be inventory in the upper oceanic water column, from the air–sea interface until the layer where ⁷Be activity is below the detection limit (in > 400 L water sample), can indirectly yield the bulk depositional flux of ⁷Be (Brost et al., 1991). This requires precise determination of the penetration depth of ⁷Be in the water column. The only uncertainty is the loss of ⁷Be-laden sinking particles from the upper water column where ⁷Be is present. After being deposited on the ocean surface, ⁷Be is generally mixed uniformly within the surface mixed layer (Young and Silker, 1980; Kadko and Olson, 1996). In open ocean, the particle concentration is generally low, and a major fraction of ⁷Be is expected to be in the dissolved phase, thus allowing particle scavenging losses to be ignored (Silker, 1972; Andrews et al., 2008). Therefore, in the absence of physical removal processes other than radioactive decay, the input flux of ⁷Be should be balanced by the ⁷Be inventory integrated over the water column. In other words, ⁷Be flux of atmospheric fallout (Bq m⁻² d⁻¹) can be obtained from the ⁷Be water column inventory (Bq m⁻²) multiplied by the decay constant (0.013 d⁻¹) of ⁷Be. This method has been proven to be reliable in open ocean due to the relatively short half-life of ⁷Be and the constancy of ⁷Be deposition over broad latitudinal bands (Young and Silker, 1980; Aaboe et al., 1981). It is expected that the ⁷Be inventory is season dependent in areas with large seasonal variations in precipitation (e.g., monsoon-dominated continental and oceanic areas). A time series study in Bermuda has shown that the inventory of ⁷Be was relatively constant throughout the year, such that ⁷Be inventory measured at any one time is likely representative (to within 20 %) of the instantaneous ⁷Be flux (Kadko and Prospeo, 2011; Kodko et al., 2015). This method is not suitable for coastal and estuarine areas where ⁷Be is scavenged substantially by particulate matter (Olsen et al., 1986; Baskaran and Santschi, 1993) and upwelling-dominated areas where ⁷Be inventory is diluted by ⁷Be “dead” water (Kadko and Johns, 2011; Haskell et al., 2015).

Another potential candidate is undisturbed soil profiles. However, ⁷Be inventories in undisturbed soils were reported to vary by more than an order of magnitude within 1 year (Walling et al., 2009; Kaste et al., 2011; Zhang et al., 2013). Such large variations in depositional fluxes of ⁷Be are attributed to the seasonal fluxes of ⁷Be. Earlier studies have shown that the atmospheric fluxes are highly dependent upon the amount of precipitation (Baskaran, 1995; Du et al., 2015). Since seasonal variations will significantly affect the ⁷Be inventory in soils, the data of ⁷Be soil inventory are not included in our dataset.

2.2.3 Indirect ²¹⁰Pb flux measurements

Several archives (soils, snow and ice cores, sediment cores, etc.) have been used to assay the ²¹⁰Pb fluxes. Here we only present ²¹⁰Pb fluxes estimated from soil profiles and snow and ice cores. The former is the most frequently used, and the latter perfectly fills the regional gap in polar regions and montane permanent snowfields. There are many ²¹⁰Pb measurements in sediment cores; however, due to the sediment focusing and erosion, most sediment cores do not provide a reliable estimate of the atmospheric ²¹⁰Pb flux (Turekian et al., 1977; Preiss et al., 1996); thus, this type of ²¹⁰Pb depositional flux data is not included in our dataset.

²¹⁰Pb in surface (upper ∼ 30 cm) soil has two sources: one is generated from the decay of ²²²Rn in the soil minerals, known as supported ²¹⁰Pb, which is produced from the decay of ²³⁸U, and the other comes from atmospheric deposition as unsupported ²¹⁰Pb. The fallout of ²¹⁰Pb is retained generally in the organic-rich surface soils, presumably because of the sequestering properties of the organo-mineral complexes (Covelo et al., 2008). When soil CO₂ combines with percolating water, carbonic acid is produced, which can leach some of the sorbed ²¹⁰Pb, ultimately resulting in slow migration down to a depth of up to 20–30 cm (Matisoff and Whiting, 2012). As a result, the surface soil layer contains excess ²¹⁰Pb compared to that from its equilibrium with ²²⁶Ra (Mabit et al., 2014). This part of the ²¹⁰Pb excess is termed “unsupported” or “excess” ²¹⁰Pb (²¹⁰Pb_ex). ²¹⁰Pb_ex is the difference between total (measured) ²¹⁰Pb and the supported ²¹⁰Pb in the soils. Supported ²¹⁰Pb is assumed to be the same as ²²⁶Ra activity, under the assumption of a secular equilibrium between ²²⁶Ra and supported ²¹⁰Pb. It can also be obtained by assuming that the supported ²¹⁰Pb activity is equal to the total ²¹⁰Pb at depths greater than 30 cm in the soil profile where atmospherically delivered ²¹⁰Pb has not reached (Matisoff, 2014). The mean residence time of ²¹⁰Pb over a large drainage basin is on the order of 2000–3000 years in surface soils (Benninger et al., 1975; Dominik et al., 1987), so the inventory of ²¹⁰Pb_ex in a soil profile that has not been disturbed by erosion, accumulation, or human activities for about a century can be used to calculate the depositional flux (Graustein and Turekian, 1986). At steady state, the ²¹⁰Pb depositional flux can be deduced using the ²¹⁰Pb_ex inventory multiplied by the decay constant (0.0311 yr⁻¹) of ²¹⁰Pb. This method has been widely used worldwide (e.g., Nozaki et al., 1978; Graustein and Turekian, 1986, 1989; Dörr and Munich, 1991; García-Orellana et al., 2006). At undisturbed soil sites, flux values derived from soil profile measurements were consistent with direct atmospheric flux observations (Olsen et al., 1985; Appleby et al., 2002, 2003), and ²¹⁰Pb_ex soil inventory showed little discrepancy at different sampling times (Porto et al., 2006, 2016).

Goldberg (1963) was the first to show that the total ²¹⁰Pb activity in a glacier from Greenland decreased with depth, with a possibility of dating ice cores. Subsequently, snow chronology in the Antarctic was determined (Crozaz et al., 1964; Picciotto et al., 1964). Since then, this technique has been used in both the large ice caps of Antarctica (e.g., Picciotto et al., 1968; Koide et al., 1979; Nijampurkar et al., 2002) and the Arctic (e.g., Crozaz and Langway, 1966; Koide et al., 1977; Dibb, 1990a, 1992; Peters et al., 1997) and small montane permanent snowfields (e.g., Windom, 1969; Gäggeler et al., 1983; Monaghan and Holdsworth, 1990). The ²¹⁰Pb flux in snow and ice cores is calculated in the same way as for the soil, except that supported ²¹⁰Pb in snow and ice cores is very low due to low concentration of ²²⁶Ra in snow and ice core and may be negligible, as the lithogenic dust is the primary source of ²²⁶Ra and its concentration in polar regions is very low (Preiss et al., 1996). When the snow accumulation rate is known, the depositional flux can also be obtained by using ²¹⁰Pb concentration in surface snow multiplied by the accumulation rate (Pourchet et al., 1997; Suzuki et al., 2004). The uncertainty in the depositional flux of ²¹⁰Pb from the snow and ice core record is the potential post-depositional movement of the snow and ice due to heavy wind and the possibility of snow melting and percolation.

Figure 1 Maps showing sampling locations of (a) ⁷Be concentration in surface air, (b) ²¹⁰Pb concentration in surface air, (c) ⁷Be atmospheric depositional flux, and (d) ²¹⁰Pb atmospheric depositional flux.

2.3 Data collection

In order to compile the global dataset for annual ⁷Be and ²¹⁰Pb air concentrations and depositional fluxes comprehensively, we attempted to collect published papers between 1955 and early 2020 in which hard data for their concentrations and depositional fluxes are available or their calculated values are reported. Using a series of keywords or with a search using a combination of words (e.g., ⁷Be, ²¹⁰Pb, air concentration, depositional flux, fallout radionuclide, atmospheric tracer, or soil erosion), data were retrieved from online literature databases (Web of Science, Science Direct, and China Knowledge Resource Integrated Database). During the literature survey, no a priori criteria (e.g., study area, sampling period, and measurement method) were applied. However, a critical review of the collected literature was conducted to obtain long-term data using the following criteria. For concentrations in air and directly measured fluxes of ⁷Be and ²¹⁰Pb, only those sites with more than 1 year of data were included. When averaged over a longer period, data are more representative because of the inherent seasonal variations (at least a full year of data) and inter-annual fluctuations (multi-year data). For indirectly measured fluxes of ⁷Be and ²¹⁰Pb, only those undisturbed sites clearly stated in the original literature were included.

Here we did not include unpublished data, as data quality control could be a potential issue. In the peer-reviewed published data, it is assumed that the authors, reviewers, and/or editors of the original articles have undertaken the necessary steps to verify data quality. All concentrations in air were converted into mBq m⁻³ and all depositional fluxes were converted into Bq m⁻² yr⁻¹ in cases where data were not already reported in these units. When available, the metadata of latitude, longitude, altitude, sampling date, annual precipitation, methodology, and references are also given. A brief description of different variables that could affect the data is summarized in Table 1. In cases where the air concentration and depositional flux data were only available graphically, a computer program (GetData Graph Digitizer) was used to digitize the data from graphics; the same was done for geographical location and annual precipitation. In rare cases, only the locality name of the study site was available, and thus the geographical coordinates were extracted from Google Earth.

Table 1 Fields in the main data table.

^a The field only appears in the ⁷Be or ²¹⁰Pb annual concentration worksheet. ^b The field only appears in the ⁷Be or ²¹⁰Pb annual depositional flux worksheet.

3 Results and discussions

3.1 Scope of the dataset

From 456 references (Appendix A), we have compiled a comprehensive dataset of atmospheric ⁷Be and ²¹⁰Pb measurements made by numerous laboratories. The dataset includes 494 annual surface air concentration data of ⁷Be covering 367 different sites, 366 annual surface air concentration data of ²¹⁰Pb from 270 different sites, 304 annual depositional flux data of ⁷Be from 279 different sites, and 645 annual depositional flux data of ²¹⁰Pb from 602 different sites. In some cases, data collected from different periods were published in different articles. In these cases, all data from the same site are listed as separate datasets; however, in data analysis and plotting, the data from the same site are merged. The sampling locations of this global dataset show broad geographical coverage (Fig. 1). The sampling maps are mainly composed of continental sites, while the oceanic monitoring sites are limited.

The number of peer-reviewed journal articles published annually containing ⁷Be and ²¹⁰Pb data from 1955 to 2020 that are included in this article is plotted in Fig. 2. Measurements of ⁷Be began in the mid-1950s (Arnold and Al-Salih, 1955; Cruikshank et al., 1956; Rama Thor and Zutshi, 1958), earlier than those of ²¹⁰Pb, which began in early 1960s (Burton and Stewart, 1960; Crozaz et al., 1964; Peirson et al., 1966). The long-term monitoring work started in the 1980s with the ⁷Be and ²¹⁰Pb concentration data generated by the EML Surface Air Sampling Program (Feely et al., 1989; Larsen et al., 1995). This was followed by more ambitious international programs, such as the REM network (Hernandez-Ceballos et al., 2015) and IMS-CTBTO (Terzi and Kalinowski, 2017). However, in these two programs, ²¹⁰Pb concentration measurements were conducted only at a few stations (Heinrich et al., 2007; Sangiorgi et al., 2019). In contrast, direct ⁷Be and ²¹⁰Pb flux measurements were rarely supported by the international program, but there were several national monitoring programs initiated by developed countries like Australia (Bonnyman and Molina-Ramos, 1971), Japan (Narazaki et al., 2003; Yamamoto et al., 2006), the United States (Lamborg et al., 2013), and Finland (Paatero et al., 2015; Leppanen, 2019). The measurement of ⁷Be inventory in the ocean mixed layer began in the 1970s (Silker, 1972; Young and Silker, 1974, 1980), and the idea proposed by Young and Silker was subsequently developed in the upper 100–200 m to assess surface water subduction, oxygen utilization, and the rate of upwelling (Kadko, 2009; Kadko and Olson, 1996; Kadko and Johns, 2011). The measurement of ²¹⁰Pb_ex in an undisturbed soil profile was first conducted by Fisenne (1968). Subsequently, Benninger et al. (1975) and Moore and Poet (1976) showed that excess ²¹⁰Pb activities in undisturbed soil profiles can be utilized to estimate the atmospheric ²¹⁰Pb depositional flux, which resulted in an increase in the measurements of ²¹⁰Pb_ex in soil profiles in the late 1980s (Graustein and Turekian, 1986, 1989; Monaghan, 1989; Dörr and Munnich, 1991). Subsequently, ²¹⁰Pb_ex was shown to be useful for soil erosion studies on agricultural land (Walling and He, 1999; Walling et al, 2003).

Figure 2 Distributions of data published between 1955 and early 2020: (a) ⁷Be concentration, (b) ²¹⁰Pb concentration, (c) ⁷Be depositional flux, and (d) ²¹⁰Pb depositional flux.

The histogram of sampling durations of ⁷Be and ²¹⁰Pb measurements is given in Fig. 3. In general, the duration of sampling for ⁷Be measurements, especially for air concentration, is longer than that of ²¹⁰Pb. Globally, there are 140 sites that monitored ⁷Be air concentration for more than 10 years. The long-term (decades) measurements of ⁷Be were mainly dedicated to investigating the effect of changes in sunspot number on ⁷Be (Megumi et al., 2000; Cannizzaro et al., 2004; Kulan et al., 2006; Pham et al., 2013; Steinmann et al., 2013). Due to the simpler measurement procedure for air concentration, the duration of air concentration measurements, whether for ⁷Be or ²¹⁰Pb, is generally longer than that of ²¹⁰Pb and/or ⁷Be depositional flux measurements.

Figure 3 Histogram of sampling durations of (a) ⁷Be concentration, (b) ²¹⁰Pb concentration, (c) ⁷Be depositional flux, and (d) ²¹⁰Pb depositional flux. For those sites with multiple datasets, we added the sampling duration together, after deducting the overlapping period (if there was one). Note that the ⁷Be and ²¹⁰Pb depositional flux data plotted here refer to those obtained using rain collectors.

Figure 4 Global distribution of (a) ⁷Be annual air concentration (mBq m⁻³), (b) ²¹⁰Pb annual air concentration (mBq m⁻³), (c) ⁷Be annual depositional flux (Bq m⁻² yr⁻¹), and (d) ²¹⁰Pb annual depositional flux (Bq m⁻² yr⁻¹). For those sites with multiple datasets, a weighted average was calculated based on sampling durations.

Figure 5 Latitudinal variability (box and whisker plots) of (a) ⁷Be concentration, (b) ²¹⁰Pb concentration, (c) ⁷Be depositional flux, and (d) ²¹⁰Pb depositional flux plotted by 10^∘ latitudinal bands.

3.2 Global variability

The global data of ⁷Be and ²¹⁰Pb air concentrations and depositional fluxes are presented in Fig. 4. The range of concentrations of ⁷Be and ²¹⁰Pb are 0.33–17.77 mBq m⁻³ and 0.003–4.65 mBq m⁻³, respectively. The range of depositional fluxes of ⁷Be and ²¹⁰Pb are 59–6350 and 1–2539 Bq m⁻² yr⁻¹, respectively. The concentrations and depositional fluxes of ⁷Be show discernable latitudinal variability (Fig. 5a and c). In general, ⁷Be concentration and flux peak at the mid-latitudes and decrease toward the Equator and poles, as was theoretically predicted by Lal and Peters (1967). A symmetric pattern is observed between the Northern Hemisphere and Southern Hemisphere, however, a sharp increase in ⁷Be air concentration (lack of flux data) occurred on the Antarctic, which reflects the subsidence of stratospheric air masses over the Antarctic continent (Wagenbach et al., 1988; Elsässer et al., 2011). Although the ²¹⁰Pb concentration and depositional flux are expected to heavily depend on the source(s) of air mass(es) and not on the latitude, the 10^∘ latitudinal variability of ²¹⁰Pb concentration and depositional flux is observed in the Northern Hemisphere and Southern Hemisphere (Fig. 5b and d). The latitudinal variability of ²¹⁰Pb flux is similar to the global fallout curve based on 167 global sites (Baskaran, 2011). Since most of the ²¹⁰Pb data are derived from continental sites, the latitudinal variation is mostly due to differences in the radon emanation rates with latitude. As the area of the landmass in the Southern Hemisphere is smaller than in the Northern Hemisphere, the ²¹⁰Pb concentration and depositional flux are much lower there. An asymmetry is observed in the ²¹⁰Pb concentration and depositional flux between the Northern Hemisphere and Southern Hemisphere, with the highest values appearing in the mid-latitudes of the Northern Hemisphere.

Figure 6 Global distribution of the fraction of dry to total depositional fluxes of (a) ⁷Be and (b) ²¹⁰Pb. (c) Box and whisker plots showing the comparison between the fraction of dry to total depositional fluxes of ⁷Be and the fraction of dry to total depositional fluxes of ²¹⁰Pb. (d) The fraction of dry to total depositional fluxes of ⁷Be versus annual precipitation. (e) The fraction of dry to total depositional fluxes of ²¹⁰Pb versus annual precipitation.

3.3 Fraction of dry deposition and effect of precipitation on ⁷Be and ²¹⁰Pb depositional flux

It was reported that dry deposition of ⁷Be and ²¹⁰Pb generally accounts for less than 10 % of the total deposition (Talbot and Andren, 1983; Brown et al., 1989; Todd et al., 1989); however, the fraction of dry deposition of ⁷Be and ²¹⁰Pb is highly variable (McNeary and Baskaran, 2003; Pham et al., 2013). It is likely that the contribution of dry fallout could increase when annual precipitation decreases (McNeary and Baskaran, 2003). The fraction of dry to total depositional flux of ⁷Be and ²¹⁰Pb are presented in Fig. 6a and b. Globally, the fraction of dry to total depositional flux of ⁷Be and ²¹⁰Pb ranged from 1 % to 44 % (mean: 12±9 %; n= 29, excluding one extreme site without precipitation) and from 5 % to 51 % (mean: 21±12 %; n= 26), respectively (Fig. 6c). The low fraction of dry to total depositional fluxes of ⁷Be and ²¹⁰Pb suggest that these nuclides are removed from the atmosphere primarily by precipitation (both rain and snowfall). Our results also support previous studies (Baskaran et al., 1993; Benitez-Nelson and Buesseler, 1999) that the fraction of dry deposition is higher for ²¹⁰Pb than for ⁷Be. The fraction of dry fallout of ⁷Be and ²¹⁰Pb is plotted against annual precipitation in Fig. 6d and e, and a weak negative correlation is observed, especially for ²¹⁰Pb.

As precipitation is the primary mechanism of removal of these nuclides from the atmosphere, the annual depositional fluxes generally depend on the amount and frequency of precipitation. In our dataset, the world's lowest ⁷Be depositional flux (only 59 Bq m⁻² yr⁻¹, less than 5 % of the global average ⁷Be flux) occurred in the Judean Desert, a precipitation-free area in the horse latitudes (Belmaker et al., 2011). The highest ⁷Be (6350 Bq m⁻² yr⁻¹) and ²¹⁰Pb (2539 Bq m⁻² yr⁻¹) depositional flux were observed in heavy rainfall areas: Hokitika, New Zealand (Harvey and Matthews, 1989), and Taiwan (Huh and Su, 2004), respectively. Positive correlations between annual depositional flux and precipitation have been observed on a local scale (e.g., Narazaki et al., 2003; García-Orellana et al., 2006; Sanchez-Cabeza et al, 2007; Leppanen et al., 2019). Here we illustrate the effect of precipitation on annual depositional flux on a global scale (Figs. 7 and 8). As both ⁷Be and ²¹⁰Pb depositional flux show latitudinal variability, the linear best-fit curve of annual depositional flux to annual precipitation is plotted within the 10^∘ latitudinal bands (if data are available). The empirical equations and fitting parameters describing the relationships between annual precipitation and ⁷Be and ²¹⁰Pb depositional fluxes are summarized in Table 2. ⁷Be and ²¹⁰Pb annual depositional fluxes generally show a good positive correlation with annual precipitation, although the data are limited in some latitudinal bands. Note that the frequency of precipitation is also an important factor that was not considered in earlier studies.

Figure 7 Annual depositional fluxes of ⁷Be are plotted against annual precipitation for 10^∘ latitudinal bands.

Figure 8 Annual depositional fluxes of ²¹⁰Pb are plotted against annual precipitation for 10^∘ latitudinal bands.

Table 2 A summary of empirical equations and fitting parameters describing the relationships between annual precipitation (x) and ⁷Be and ²¹⁰Pb annual depositional fluxes (y) for different latitudinal bands.

3.4 ⁷Be $/$ ²¹⁰Pb ratios and deposition velocities

Considering that some data come from the same station, we further calculated the ratios of ⁷Be to ²¹⁰Pb and deposition velocities of aerosols using ⁷Be and ²¹⁰Pb data, as shown in Figs. 9 and 10.

The variations in the ⁷Be $/$ ²¹⁰Pb ratios reflect both vertical and horizontal transport in the atmosphere (Baskaran, 1995; Koch et al., 1996; Arimoto et al., 1999; Lee et al., 2007; Tositti et al., 2014). Our dataset exhibits similar global patterns of ⁷Be $/$ ²¹⁰Pb ratio as have been simulated with a three-dimensional chemical tracer model (Koch et al., 1996), with a positive south poleward gradient and a little variation in the Northern Hemisphere. Globally, the ⁷Be $/$ ²¹⁰Pb air concentration ratio ranged from 2 to 222, and the ⁷Be $/$ ²¹⁰Pb depositional flux ratio ranged from 2 to 229. At the 19 sites for which ⁷Be $/$ ²¹⁰Pb air concentration ratio and depositional flux ratio were available simultaneously, a paired t test indicates that at 0.05 level the ⁷Be $/$ ²¹⁰Pb air concentration ratio and depositional flux ratio are not significantly different.

Figure 9 Global distribution of (a) ⁷Be $/$ ²¹⁰Pb activity ratio and (b) ⁷Be $/$ ²¹⁰Pb depositional flux ratio. Note that the color bar below applies to both figures.

⁷Be and ²¹⁰Pb are excellent tracers for the determination of the deposition velocities of aerosols for two reasons: (1) their depositional fluxes and air concentrations at any given site remain fairly constant over a long period and (2) their size distributions in aerosols are similar to that of many particulate contaminants of interest (McNeary and Baskaran, 2003; Dueñas et al., 2005). When both air concentration (C) and depositional flux (F) at the same site are available, the average total deposition velocities of aerosols that carry these nuclides (V_d) can be calculated by the following Eq. (1):

$$\begin{array}{}\text{(1)}& V_{\text{d}}=F/C.\end{array}$$

Thus, the V_d obtained from ⁷Be and ²¹⁰Pb can be used to determine the depositional flux of analog species with a knowledge of their air concentration (Turekian et al., 1983). The V_d for ⁷Be ranged from 0.2 to 8.4 cm s⁻¹ (mean: 1.3±1.2 cm s⁻¹; n= 70), and for ²¹⁰Pb it ranged from 0.1 to 12.7 cm s⁻¹ (mean: 1.2±1.7 cm s⁻¹; n= 72). The deposition velocity of aerosols collected over a period of 17 months in Detroit, MI, USA, varied over 2 orders of magnitude, from 0.2 to 3.6 cm s⁻¹ (mean: 1.6 cm s⁻¹; n= 30) for ⁷Be and 0.04 to 3.6 cm s⁻¹ (mean: 1.1 cm s⁻¹; n= 30) (McNeary and Baskaran, 2003). A summary of deposition velocity from 10 different stations is also given in McNeary and Baskaran (2003). Earlier studies suggested that at continental sites V_d of ⁷Be will be higher than V_d of ²¹⁰Pb using the ground level as the reference, which is an artifact in the manner in which the calculation is made (Turekian et al., 1983; Todd et al., 1989; McNeary and Baskaran, 2003). However, later works observed opposite results (Dueñas et al., 2005, 2017; Lozano et al., 2011; Mohan et al., 2019). The independent t test analysis indicates that at the 0.05 level the V_d calculated by ⁷Be and ²¹⁰Pb are not significantly different in the global dataset, which suggests that ⁷Be and ²¹⁰Pb attach onto the aerosols by similar mechanisms (Winkler et al., 1998; Papastefanou, 2006) and are affected by similar deposition processes (Lozano et al., 2013).

Figure 10 Global distribution of deposition velocities (V_d) of aerosols (cm s⁻¹) obtained from (a) ⁷Be and (b) ²¹⁰Pb. Note that the color bar below applies to both figures.

3.5 Investigations of global atmospheric dynamics and climate changes

Developing numerical models in which aerosols, chemistry, radiation, and clouds interact with one another and with atmospheric dynamics is important for understanding and predicting global climate changes (Brost et al., 1991). In such atmospheric dynamic models, the major uncertainty is from the parameterization of subgrid-scale processes such as precipitation scavenging, vertical transport, and radiative effect. Taken together, the cosmogenic ⁷Be and terrigenous ²¹⁰Pb offer an excellent tool in investigating wet scavenging and vertical transport in global models (Liu et al., 2001).

A set of data obtained prior to the 1990s was used to compare simulated results in global models such as ECHAM2 (Brost et al., 1991; Feichter et al., 1991), ECMWF (Rehfeld and Heimann, 1995), chemical transport model (CTM) based on GISS GCM (Balkanski et al., 1993; Koch et al., 1996), GEOS-CHEM (Liu et al., 2001), LMDz (Preiss and Genthon, 1997; Heinrich and Pilon, 2013), and GMI CTM (Liu et al., 2016). By simulating the ratio of ⁷Be $/$ ²¹⁰Pb, Koch et al. (1996) eliminated the error associated with the effect of precipitation and provided a better measure of vertical transport. After correcting the cross-tropopause transport, simulation of observed ⁷Be and ²¹⁰Pb surface concentrations and depositional fluxes with no significant global bias was obtained (Liu et al., 2001). Note that the spatial coverage of the dataset used in the previous modeling work was only partial and thus limited the statistical significance of comparisons of simulated and observed results (Feichter et al., 1991). Additional work with more data is needed for detailed comparison and successful validation of models (Brost et al., 1991). The size of the ⁷Be and ²¹⁰Pb datasets has greatly increased in the last 3 decades, and our new dataset is expected to lay a foundation for developing better parameterizations and to contribute to modeling efforts.

3.6 Soil erosion, aquatic particle dynamics, and ocean surface process studies

The atmospheric depositional flux data of ⁷Be and ²¹⁰Pb are useful in utilizing these nuclides as tracers for soil erosion and redistribution studies in terrestrial environments (Mabit et al., 2008) and particle dynamics studies in aquatic environments (e.g., Du et al., 2012; Matisoff, 2014). The basic principle involved in using ⁷Be or ²¹⁰Pb_ex as soil tracers are the same, which is to compare the measured inventory of ⁷Be or ²¹⁰Pb_ex (Bq m⁻²) at a sampling point with the inventory at an undisturbed (or reference) site (Blake et al., 1999; Walling and He, 1999). Depletion of the inventory means that soil erosion has occurred, whereas enrichment provides evidence of accumulation of surficial soil. The first step in these studies is to select a suitable undisturbed site and obtain the reference inventory (Mabit et al., 2009). However, as human activity intensifies, such undisturbed sites are not always readily available. In aquatic systems (including rivers, lakes, estuaries, and coasts), the mass balance models of ⁷Be and ²¹⁰Pb_ex have become powerful tools to understand the sediment source, transportation and resuspension processes (e.g., Wieland et al., 1991; Feng et al., 1999; Jweda et al., 2008; Huang et al., 2013; Mudbidre et al., 2014). In such models, the atmospheric depositional input of ⁷Be and ²¹⁰Pb is a required source term. Furthermore, the ⁷Be $/$ ²¹⁰Pb_ex activity ratio can be used to identify the source area of sediments (Whiting et al., 2005; Jweda et al., 2008; Wang et al., 2021), to quantify the age of sediments (Matisoff et al., 2005; Saari et al., 2010), and to determine the transport distance of suspended particles (Bonniwell et al., 1999; Matisoff et al., 2002). Thus, the atmospheric depositional flux data of ⁷Be and ²¹⁰Pb are also important for tracing particle dynamics in aquatic systems.

The atmospheric depositional flux (or ocean inventory) data of ⁷Be serve as an indispensable parameter for tracing surface ocean process (e.g., subduction, upwelling, and depositional flux of trace metals) (Kadko, 2017; Kadko and Olsen, 1996; Kadko et al., 2015). Due to the low activity of ⁷Be in open-ocean waters, usually 400–700 L of seawater is needed, which imposes some limitations for sampling, especially for deep layers. This constraint has limited its application. If the ⁷Be ocean inventory can be accurately estimated, the collection of a large volume of seawater can be avoided and the application of ⁷Be in open ocean will be expanded.

Scientific data are not simply the outputs of research, but they also provide inputs to new hypotheses, extending research and enabling new scientific insights (Tenopir et al., 2011). Our dataset provides a forum in which a large amount of ⁷Be and ²¹⁰Pb atmospheric depositional flux data for the above-mentioned research communities. This database will help in identifying data gaps and evaluating the empirical relations between ⁷Be and ²¹⁰Pb depositional fluxes and annual precipitation (Table 2). Researchers can rely on previously collected data in planning their research, without additional monitoring of ⁷Be and/or ²¹⁰Pb depositional fluxes. Even for those areas with data gaps, the empirical equations between ⁷Be and ²¹⁰Pb depositional fluxes and annual precipitation provide an empirical method for estimating fluxes, especially for ⁷Be, as ⁷Be depositional flux is independent of longitude and is constant over broad latitudinal bands. In summary, the atmospheric depositional flux data presented in this dataset, along with the meta-analysis of the data, will be useful in the investigations of soil erosion studies in terrestrial environments, particle dynamics studies in aquatic systems, and surface mixing process studies in open ocean.

3.7 Gaps and recommendations

Our dataset is a compilation of most of the published results in international peer-reviewed journal articles. Although the spatial coverage of this dataset is significant, the available sites are unevenly distributed. Compared to spatial and temporal coverage of depositional flux data, the coverage of air concentration data is much larger. The air concentration measurements from deep-ocean sites are limited, but the data from coastal oceanic sites are adequate; however, the depositional flux measurement at oceanic sites is rare. Concerning air concentrations, areas such as Europe, East Asia, eastern Oceania, and the eastern United States are well covered, whereas other areas such as the African continent and northern Asia are limited. A similar spatial coverage pattern exists for the depositional flux of these nuclides, but the regional gaps are more notable, especially for ⁷Be flux data, which are almost non-existent for the Antarctic and African continents. In addition, it needs to be emphasized that the number of sampling sites, in which both concentration and flux of ⁷Be and ²¹⁰Pb were measured simultaneously, are limited.

We recommend that future studies pay more attention to those areas that are currently undersampled or unsampled to better characterize the expected global variability in the ⁷Be and ²¹⁰Pb air concentrations and depositional fluxes by measuring both nuclides simultaneously to obtain more data, as well as the ⁷Be $/$ ²¹⁰Pb ratio, and estimate the deposition velocity of aerosols. In areas with very limited precipitation, such as deserts, it is expected that the dry fallout will dominate the bulk depositional flux, and quantification of the role of dry fallout in the removal of these nuclides will provide insights into the removal of other analog species. As mentioned earlier, combining cosmogenic ⁷Be with ²¹⁰Pb that predominantly originates from the Earth's surface will be useful to trace species that originate from the Earth's surface, such as Hg, SO${}_{\mathrm{4}}^{\mathrm{2}-}$, NO${}_{\mathrm{3}}^{-}$, and those that originate in the upper atmosphere, such as O₃. The size distribution of aerosol particles carrying ⁷Be and ²¹⁰Pb is crucial for understanding atmospheric behavior and tracing analogues, and such studies also need to be conducted. Besides, the troposphere contains ∼ 99 % of global water vapor with < 1 % in the stratosphere. The deposition velocity of aerosol in the stratosphere is very low (∼ $\mathrm{4}\times {\mathrm{10}}^{-\mathrm{3}}$ cm s⁻¹; Junge, 1963) with no precipitation. Thus, the ⁷Be concentration is governed by local production, zonal and vertical downward transport, and its decay. In the middle and upper troposphere where precipitation is much less frequent, the removal rate of aerosols is also slow. Collection of air samples in that part of the atmosphere will provide useful information on the total deposition velocity of aerosols (Lal and Baskaran, 2012).

Finally, we acknowledge that the seasonal data of ⁷Be and ²¹⁰Pb have not been included in the current version of dataset because compiling the seasonal data is more challenging than compiling the annual data. Unlike the annual data, most of the published seasonal data are presented in graphs, without being given in tables, and in some cases the graph quality was poor, and thus the precision of the data extraction is expected to be poor. Besides, in some papers, although seasonal data were measured, only the annual data were provided. Thus, the comprehensive compilation of seasonal data of ⁷Be and ²¹⁰Pb may need collaboration with, and data sharing from, the scientific community. The compilation of seasonal data is expected to be useful to assess seasonal variability of ⁷Be and ²¹⁰Pb and understand the relationship between these changes and influencing factors, such as atmospheric dynamics, meteorological conditions, and geographic location, on a global scale. The seasonal data can also be useful in evaluating the seasonality of transport in global atmospheric models.

4 Data formats and availability

For clarity and convenience, four separate worksheets, each named as ⁷Be or ²¹⁰Pb annual air concentration and ⁷Be or ²¹⁰Pb annual atmospheric flux, are available in one Microsoft Excel^® file, although sometimes these data come from the same article. In addition, the data of the ⁷Be $/$ ²¹⁰Pb air concentration ratio, ⁷Be $/$ ²¹⁰Pb depositional flux ratio, and deposition velocities for ⁷Be and ²¹⁰Pb are also presented. The dataset can be downloaded from Zenodo (https://doi.org/10.5281/zenodo.4785136, Zhang et al., 2021). It is free for scientific applications, but the free availability does not constitute a license to reproduce or publish it. The compilation of seasonal data is ongoing, and the dataset will be updated in a future effort.

5 Conclusions

This paper summarizes the global dataset of ⁷Be and ²¹⁰Pb for their concentration in atmospheric air and their depositional fluxes from 456 publications spanning the period from 1955 to early 2020. The calculated activity ratios of ⁷Be $/$ ²¹⁰Pb and deposition velocity of aerosols are also reported. Some noteworthy spatial gaps in the dataset are the African continent, northern Asia, and Antarctica (only for ⁷Be flux). Despite these gaps, our dataset is the largest compilation of ⁷Be and ²¹⁰Pb air concentration and depositional flux up to date and could be used to better understand the transport processes of air masses and depositional processes of aerosols. This dataset not only lays a solid foundation to develop better parameterizations contributing to future modeling efforts but also supplies a basic parameter for tracing soil erosion on land, particle dynamics in aquatic systems, and ocean surface processes using ⁷Be and/or ²¹⁰Pb.

Appendix A: List of references in the dataset

Aaboe et al. (1981), Aba et al. (2016), Ahmed et al. (2004), Akata et al. (2008, 2015, 2018a, 2018b), Akram et al. (1999), Al-Azmi et al. (2001), Alegría et al. (2010), Ali et al. (2011a, b), Alonso-Hernández et al. (2004, 2014), Amano and Kasai (1981), Anand and Rangarajan (1990), Anderson et al. (1960), Andres (2018), Appleby et al. (2002, 2003, 2019), Arimoto et al. (1999), Arkian et al. (2010), Arnold and Al-Salih (1955), Azahra et al. (2003, 2004), Azimov et al. (2011, 2017), Bachhuber and Bunzl (1992), Baeza et al. (1996, 2016), Bas et al. (2017), Baskaran and Swarzenski (2007), Baskaran et al. (1993), Batraov et al. (2016), Bazarbaev et al. (2012), Begy et al. (2016), Beks et al. (1998), Belmaker et al. (2011), Belyaev et al. (2004), Benitez-Nelson and Buesseler (1999), Benmansour et al. (2013), Bettoli et al. (1995), Bikkina et al. (2015), Blake et al. (2009), Blazej and Mietelski (2014), Bleichrodt (1978), Bleichrodt and van Abkoude (1963), Bourcier et al. (2011), Brandt et al. (2018), Branford et al. (2004), Brost et al. (1991), Brown et al. (1989), Buck et al. (2019), Buraeva et al. (2013a, 2013b), Burton and Stewart (1960), Caillet et al. (2001), Cámara-Mor et al. (2011), Cannizzaro et al. (1999, 2004), Canuel et al. (1990), Cao et al. (2018), Carpenter et al. (1981), Carvalho et al. (1995, 2013), Chae and Kim (2019), Chae et al. (2011), Chang et al. (2008), Chao et al. (2012, 2014), Chen (2014), Chen et al. (2016, 2020), Chham et al. (2017, 2018, 2019), Cho et al. (2011), Clifton et al. (1995), Conaway et al. (2013), Courtier et al. (2017), Crecelius (1981), Crozaz and Langway (1966), Crozaz et al. (1964), Cruikshank et al. (1956), Cruz et al. (2019), Cui et al. (2012), Daish et al. (2005), Damatto et al. (2005), Damnati et al. (2013), D'Amours et al. (2013), de Tombeur et al. (2020), Deng et al. (2020), Dibb (1989, 1990a, 1990b, 1992, 2007), Dibb and Jaffrezo (1993), Dibb et al. (1994), Ding et al. (2017), Dlugosz-Lisiecka (2019), Doering and Akber (2008a, 2008b), Doering and Saey (2014), Doering et al. (2006), Doi et al. (2007), Dominik et al. (1987, 1989), Dörr and Münnich (1991), Dovhyi et al. (2017), Du et al. (2008, 2015), Du and Walling (2012), Dueñas et al. (1999, 2004, 2005, 2009, 2011, 2017), Ďurana et al. (1996), Dutkiewicz and Husain (1985), El-Hussein et al. (2001), Elsässer et al. (2011), Eriksson et al. (2004), Fan et al. (2013), Fang et al. (2013), Feely et al. (1989), Filizok and Ugur Gorgun (2019), Filizok et al. (2013), Fogh et al. (1999), Fukuyama et al. (2008), Fuller and Hammond (1983), Gäggeler et al. (1983, 1995), Gai et al. (2015), García-Orellana et al. (2006), Garimella et al. (2003), Garspar et al. (2013), Gavini et al. (1974), Gerasopoulos et al. (2001), Gonzalez-Gomez et al. (2006), Gordo et al. (2015), Grabowska et al. (2003), Graham et al. (2003), Graustein and Turekian (1986, 1989), Grossi et al. (2016), Gustafson et al. (1961), Halstead et al. (2000), Harvey and Matthews (1989), Hasebe et al. (1981), Hasegawa et al. (2007), Haskell et al. (2015), He and Walling (1997), He et al. (2018), Heikkilä et al. (2008), Heinrich et al. (2007), Hernández et al. (2005, 2007, 2008), Hernandez-Ceballos et al. (2015), Hicks and Goodman (1977), Hirose et al. (2004), Hötzl and Winkler (1987), Houali et al. (2019), Hu (2016), Hu and Zhang (2019), Hu et al. (2020), Huang et al. (2019), Huh and Su (2004), Huh et al. (2006), Igarashi et al. (1998), Ioannidou and Paatero (2014), Ioannidou and Papastefanou (2006), Ioannidou et al. (2005, 2019), Irlweck et al. (1997), Isakar et al. (2016), Ishikawa et al. (1995), Itoh and Narazaki (2017), Itthipoonthanakorn et al. (2019), Iurian et al. (2013), Jankovic et al. (2014), Jasiulionis and Wershofen (2005), Jia and Jia (2014), Jia et al. (2003), Jiang (1999), Joshi (1985), Joshi et al. (1969), Juri Ayub et al. (2009), Kadko (2000), Kadko and Johns (2011), Kadko and Olson (1996), Kadko and Prospero (2011), Kadko and Swart (2004), Kadko et al. (2015, 2016), Kapala et al. (2018), Karwan et al. (2016), Kato et al. (2010), Khan et al. (2008, 2009), Khodadadi et al. (2018), Kikuchi et al. (2009), Kim et al. (1998, 1999, 2000, 2005), Kitto et al. (2005, 2006), Klaminder et al. (2006), Koide et al. (1977, 1979), Kolb (1970), Kownacka et al. (1990), Krmar et al. (2015), Kulan et al. (2006), Kurata and Taunogai (1986), Laguionie et al. (2014), Lal et al. (1979), Lambert et al. (1990), Lamborg et al. (2000, 2013), Landis et al. (2014), Larsen et al. (1995), Lee et al. (1985, 2002, 2015), Leppanen (2019), Le Roux et al. (2008), Li et al. (2009, 2013, 2017a, b), Likuku (2006), Likuku et al. (2006), Lin et al. (2014), Lindblom (1969), Liu et al. (2014), Lockhart et al. (1966), Lozano et al. (2011, 2012, 2013), Lujanienë (2003), Luyanas et al. (1970), Mabit et al. (2009), Maenhaut et al. (1979), Magno et al. (1970), Marx et al. (2005), Mattsson (1970), McNeary and Baskaran (2003), Megumi et al. (2000), Mélières et al. (2003), Men et al. (2016), Meusburger et al. (2016, 2018), Mietelski et al. (2017), Milton et al. (2001), Miralles et al. (2004), Mohan et al. (2018, 2019), Mohery et al. (2014, 2016), Momoshima et al. (2006), Monaghan (1989), Monaghan and Holdsworth (1990), Monaghan et al. (1986), Moore and Poet (1976), Muramatsu et al. (2008), Narazaki and Fujitaka (2009), Narazaki et al. (2003), Neroda et al. (2016), Nijampurkar and Clausen (1990), Nijampurkar and Rao (1993), Nijampurkar et al. (2002), Noithong et al. (2019), Nozaki et al. (1978), O'Farrell et al. (2007), Olsen et al. (1985), Othman et al. (1998), Paatero and Hatakka (2000), Paatero et al. (2003, 2010, 2015, 2017), Pacini et al. (2011, 2015), Padilla et al. (2019), Pan et al. (2011, 2017), Papastefanou and Bondietti (1991), Papastefanou and Ioannidou (1991), Papastefanou et al. (1995), Parker (1962), Peirson (1963), Peirson et al. (1966), Peng et al. (2019), Perreault et al. (2017), Persson (2015), Peters et al. (1997), Pfahler et al. (2015), Pfitzner et al. (2004), Pham et al. (2011, 2013), Picciotto et al. (1964, 1968), Piñero-García and Ferro-García (2013), Piñero-García et al. (2012, 2015), Poet et al. (1972), Poreba et al. (2019), Porto and Walling (2012), Porto et al. (2006, 2009, 2013, 2014, 2016), Pourchet et al. (1997), Preiss et al. (1996), Prospero et al. (1995), Qian et al. (1985), Rabesiranana et al. (2016), Rajačić et al. (2015, 2016), Raksawong et al. (2017), Ram and Sarin (2012), Rangarajan et al. (1966, 1975, 1986), Rastogi and Sarin (2008), Realo et al. (2004, 2007), Reiter et al. (1983), Renfro et al. (2013), Rodas Ceballos et al. (2016), Ródenas et al. (1997), Rodriguez-Perulero et al. (2019), Saari et al. (2010), Sabuti and Mohamed (2016), Sakurai et al. (2005, 2011), Saleh and Abdel-Halim (2017), Sambayev et al. (2019), Samolov et al. (2014), San Miguel et al. (2019), Sanchez-Cabeza et al. (2007), Sanders et al. (2011), Sato et al. (1994, 2003), Savva et al. (2018), Schuler et al. (1991), Schumann and Stoeppler (1963), Shapiro and Forbes-Resha (1976), Sheets and Lawrence (1999), Shelley et al. (2017), Shi et al. (2011, 2017), Shleien and Friend (1966), Short et al. (2007), Silker (1972), Simon et al. (2009), Smith et al. (1997), Song et al. (2003, 2015), Stamoulis et al. (2018), Steinmann et al. (1999, 2013), Stromsoe et al. (2016), Su et al. (2003), Sugihara et al. (2000), Suzuki and Shiono (1995), Suzuki et al. (1999, 2004, 2017), Sykora et al. (2017), Talbot and Andren (1983), Tan et al. (2013, 2016), Tanahara et al. (2014), Tanaka and Turekian (1995), Tateda and Iwao (2008), Taylor et al. (2016), Terzi and Kalinowski (2017), Thang et al. (2018), Thompson et al. (1984), Thor and Zutshi (1958), Todd et al. (1989), Todorovic et al. (1999, 2000, 2005, 2010), Tokieda et al. (1996), Tositti et al. (2014), Tsunogai et al. (1985, 1988), Tuo et al. (2018), Turekian and Cochran (1981), Turekian et al. (1977, 1983), Uchida et al. (2009), Uematsu et al. (1994), Ueno et al. (2003), Uğur et al. (2011), Uhlář et al. (2014), Valles et al. (2009), Van Metre and Fuller (2009), Vecchi and Valli (1997), Vecchi et al. (2005), Vogler et al. (1996), Von Gunten and Moser, (1993), Wagenbach et al. (1988), Wakiyama et al. (2010), Wallbrink and Murray (1994, 1996), Walling and He (1999), Walling et al. (2003), Walton and Fried (1962), Wan et al. (2010), Wang (2010, 2011), B. Wang et al. (2014), Z. Wang et al. (2014), Weiss and Naidu (1986), Wells et al. (2012), Windom (1969), Winkler and Rosner (2000), Winkler et al. (1998), Wu et al. (2011), Yamagata et al. (2019), Yamamoto et al. (2006), Yang et al. (1999, 2011, 2013), Yi et al. (2005, 2007), Yoshimori (2005), Young and Silker (1980), Yu et al. (2017, 2018), Zanis et al. (1999, 2003), Zhang et al. (2003, 2006, 2014, 2016, 2018, 2019), Zhang and Jiang (2018), Zheng et al. (2005, 2007), and Zhu and Olsen (2009).