The composition and transport of organic carbon in rainfall: Insights from the natural (13C and 14C) isotopes of carbon



[1] The dilute nature and compositional complexity of rainwater organic carbon (OC) has made it difficult to determine the role of rainwater OC in regional and global carbon budgets. Here I present the δ13C and Δ14C signatures of OC in rainfall, including two hurricanes, which indicate that OC in rain originates from multiple sources including a significant fraction from fossil fuels and marine OC. These samples demonstrate that atmospheric wet deposition of OC from fossil fuels can supply significant quantities of relic carbon to ecosystems. In addition, rainfall events that originate over the ocean can be dominated by marine OC, arguing for a significant transfer of marine OC to the continents and a large recycled marine OC component to marine rain. A conceptual model of the direct transfers of OC in rainfall concludes that the net transfer of continental OC to the ocean could be close to zero due to a significant marine to continental flux.

1. Introduction

[2] The sources and composition of organic carbon (OC) in rainwater are still relatively unknown, leading to uncertainty in the role of rainwater OC in regional and global C budgets. For instance, mass balance studies of oceanic carbon budgets have begun to incorporate a significant input of allochthonous continental organic carbon into the oceans via atmospheric wet deposition [del Giorgio and Duarte, 2002; Hansell et al., 2004; Vlahos et al., 2002; Willey et al., 2000]. Interestingly, the net flux of water (continental to ocean water transfers minus ocean to continental transfers) is negative, or in the opposite direction of the proposed carbon flux in precipitation [Baumgartner and Reichel, 1975] and if marine precipitation is dominated by marine organic carbon [Duce and Duursma, 1977] then there could be appreciable transfers of marine OC between marine ecosystems and to the continents. Additionally, independent studies on tropospheric ozone, cloud formation, air visibility, and contamination have demonstrated the presence of organic compounds originated from fossil fuel burning in the atmosphere and in rainwater [Ducret and Cachier, 1992; Klouda et al., 1996; Lewis et al., 2004; Reddy et al., 2002; Takase et al., 2003]. The transfer of fossil fuel carbon represents a mobilization of organic carbon that was previously stored for millennia, yet the relative importance of wet deposition for relocating fossil fuel carbon to terrestrial and oceanic reservoirs is unknown. What is lacking is detailed knowledge about the transfers of OC through precipitation. Thus, studies on the quantities, sources and reactivity of organic matter in rainfall is a critical aspect of understanding the role of the wet deposition of OC to ecosystem energy budgets, yet have been hampered by the low concentrations and chemical complexity of rainwater OC [Seitzinger et al., 2003].

[3] In order to assess the quantities and sources of rainwater OC, rainwater was collected from two sites in the Northeastern United States and analyzed for 13C and 14C. A majority of the samples were taken from New Haven, CT a small city on the Long Island Sound, while one sample was taken from Woods Hole MA. Glass beakers that were pre-baked at 550°C were placed on the rooftop of the Yale Environmental Research center within 6 hours preceding a rainstorm. The beakers were then collected within 6 hours of a storm's termination and the rainwater was transferred to acid washed polycarbonate bottles and immediately frozen. Samples were not filtered for fear of contamination, but other studies have found that the large majority (∼98%) of OC in unfiltered rainwater is in the dissolved (<0.7μm) form [Willey et al., 2000]. For isotopic analysis, 300ml of sample was transferred to a baked quartz reaction tube. The sample was acidified to pH 2 with 50% phosphoric acid and sparged with UHP N2 to remove any inorganic carbon. The remaining OC was then oxidized using a high energy UV lamp in the presence of O2. The resulting CO2 was then cryogenically purified on a vacuum extraction line and sent for isotopic analysis at the NOSAMS at Woods Hole, or the University of Arizona's AMS facility.

2. Results/Discussion

[4] Isotopic signatures indicate that the OC in rainwater from the two sites does not simply originate from modern terrestrial sources (Table 1). Multiple sources of OC including a δ13C enriched and a Δ14C depleted relic contribution (Table 1) are apparent. Previous studies of the organic content of atmospheric particulate organic carbon and volatile organic compounds (VOC) have generally reported two major sources of organic carbon to the atmosphere: compounds from contemporary vegetation emissions [Guenther et al., 1995] and relic organic carbon from fossil fuel burning [Ducret and Cachier, 1992; Klouda et al., 1996; Lewis et al., 2004; Reddy et al., 2002; Takase et al., 2003]. In fact, previous isotopic studies on VOC's and polycyclic aromatic hydrocarbons (PAH's) have argued that these two sources dominate radiocarbon measurements [Lewis et al., 2004; Reddy et al., 2002]. Plant VOC emissions are difficult to measure, but are thought to be ∼0–5% of plant NPP [Chapin et al., 2002]. Fossil fuel carbon is generated from the incomplete combustion of fossil fuels and are present as volatile hydrocarbon gases and sub-micron particles which can dissolve in rain [Ducret and Cachier, 1992].

Table 1. Rainfall Amounts, Concentrations, Isotopic Ratios, and Relative Contributions of the Three Major OC Sources to Rainfall Events Estimated Using a Dual Isotopic Mixing Modela
DateRain, InOC, μMδ13C, ‰Δ14C, ‰14C-age% Marine% Fossil Fuel% Modern Continent.
  • a

    See auxiliary material.

  • b

    Collected from Woods Hole, MA; all other samples from New Haven CT.

03/28/01b 58−25.0−4414657454510
AVERAGE 52−24.8−1802122362243

[5] These two sources, however, may not explain the δ13C enriched samples (Table 1, 4/12/04, 9/17/04, 9/28/04). Marine biota produce significant quantities of volatile organic carbon that contribute to atmospheric organic carbon pools [O'Dowd et al., 2004], with another potential source of marine organic matter to rainwater and marine aerosols being marine organic matter associated with sea-spray [Duce and Duursma, 1977]. Other coastal and open ocean studies have demonstrated that rainwater originating over the ocean and sampled prior to or immediately upon contact with land contains a small yet appreciable amount of organic carbon seemingly from marine sources [Avery et al., 2004]. Marine OC is δ13C enriched and I argue that the 3 dominant sources of organic carbon to the rain samples collected by this study are modern terrestrial vegetation, fossil fuels, and marine OC sources. This 3-source contribution to rainwater OC is corroborated by the isotopic data when plotted on a source diagram (Figure 1). Using the δ13C and Δ14C values, a three-source isotopic mixing equation (see auxiliary material) is used to determine the relative contributions of these three presumed sources to rainwater OC (Table 1). This exercise demonstrates the relative importance of marine and fossil OC to rainwater collected by this study (Table 1). The four events that tracked almost exclusively over the continental U.S. are dominated by OC from fossil fuel and modern terrestrial sources (Table 1 and Figure A1), although a small contribution from marine sources cannot be ruled out using the 3 end-member mixing model. The three events that are dominated by marine OC (Table 1) all tracked over either the Gulf of Mexico or the Atlantic Ocean (Table 1 and Figure A1), which is further evidence of marine OC contributing to the δ13C enriched ratios of these events, although small contributions to these samples from fossil fuel and modern continental sources can not be ruled out.

Figure 1.

An isotopic source diagram for all rain samples.

[6] Due to its small global pool size the transfer of fossil fuel OC through wet deposition could be a considerable component of the relic carbon budget. With respect to relic OC transfers, studies (both marine and terrestrial) are now finding that large 14C-depleted OC reservoirs (e.g., soils and oceanic dissolved OC) are not dominated by a single pool with ages proximate to bulk ages but consist of a large modern pool of organic carbon originating from contemporary primary production and a small 14C-depleted relic pool that ‘ages’ the reservoir [Loh et al., 2004; Trumbore, 2000]. The ramification is that these reservoirs now demand input fluxes that are much larger than previously calculated to maintain the large modern pool, and that a small yet significant pool of relic organic carbon is sequestered on geological time scales. Hence determining the controls on the input, removals and transfers of relic carbon budget is an important component of carbon biogeochemistry.

[7] The three end-member calculation estimates that ∼20–30% of rainwater OC is relic carbon. Assuming an annual rainfall input of ∼2g of C m−2 yr−2 of rainfall OC to terrestrial systems of the Northeast US [Willey et al., 2000], results in the deposition of ∼0.5 grams of relic OC m−2 yr−1. The average flux of relic DOC from rivers to the coastal ocean from this region is ∼0.22 g relic carbon m−2 of watershed yr−1 [Raymond et al., 2004]. Thus the input of relic carbon via rainfall to watersheds is approximate to the export of watershed relic carbon to rivers. Assuming that a similar input of relic carbon occurs over other areas of the globe that are severely impacted by N deposition (∼4% of the surface area of the earth receives >1.0g N m−2 yr−1 [Galloway et al., 2004]) provides a total input of relic carbon of ∼10Tg of fossil carbon yr−1, which is ∼40% of the annual emissions of black carbon [Penner et al., 1993]. Urban centers like New Haven are numerous in the mid-west and eastern United States so as storms track across these regions and continue out to sea they also represent a direct source of relic OC to the coastal waters of the Atlantic and could explain the surface “lens” of relic OC found in ocean studies [Druffel et al., 1992]. Many other coastal regions of the world are directly adjacent to and/or downwind from highly populated urbanized areas that rely on fossil fuels, therefore the input of relic carbon to the oceans is not limited to the coastal Atlantic.

[8] The relative importance of the marine OC transfer to land is significant to local, regional and even global carbon budgets. Of the three samples dominated by marine OC, two of them were from hurricanes originating over the ocean, and all three of these marine OC dominated rainfall events moved from the Gulf Coast to the northeast U.S. before depositing rain in the rain collectors (Figure A1). Thus, despite the fact that all three of these marine OC dominated storms traveled >2000km over land, they all had intact marine OC signatures and presumably deposited marine OC over the entire storm track. Coastal studies have also noted the deposition of rainwater with low concentrations of OC on coastal sites during hurricanes and ocean originating storms [Avery et al., 2004; Kieber et al., 2002], and these data demonstrate that this material can be almost entirely marine OC. Furthermore, these storms have the ability to move marine OC far inland. An estimated ∼4.7 × 1010 grams of OC is projected to be deposited on the continental US during the transit of each hurricane from the Gulf Coast to New England (assuming a distance traveled of 2200km, storm width of 700km, average rainfall of 10cm, and OC concentration of 25μM) which is approximately equal to ∼10% of the annual DOC flux to the middle Atlantic Bight of the United States [Raymond and Bauer, 2000].

[9] Regional and global carbon budgets have recently begun to assess the potential importance of the transport of terrestrial organic carbon in rainwater as a source of reduced carbon to the oceans [del Giorgio and Duarte, 2002; Hansell et al., 2004; Vlahos et al., 2002]. However, these studies have generally not incorporated a simultaneous marine-land flux because of the paucity of data on rainfall OC. The data presented here indicate that OC in storms originating over the ocean can be dominated by marine OC and therefore a large percentage of OC in marine precipitation is arguably recycled marine OC, which highlights a mechanism for transferring marine OC between marine ecosystems and future research should determine if specific marine ecosystems receive a subsidy of wet deposition OC from neighboring or even distant marine systems. In order to provide new estimates of the transfers of OC in rainfall, insights from this work were coupled with a conceptual estimation of the net and gross continent-ocean transfers of water (Figure 2). The major assumption with this model is that the gross continental to ocean and ocean to continental water transfers are accompanied by OC fluxes in proportion to the concentration and source apportionments obtained by this study (see Figure 2 caption). This exercise provides a gross continent to ocean OC flux of ∼32 Tg (Figure 2) and a gross marine to continent OC flux of ∼25 Tg (Figure 2) and therefore a net continental to marine flux of only ∼7Tg. These estimates are first order because authors have argued for the transport of continental OM rain precursors over large oceanic distances [Chesselet et al., 1981; Church et al., 1991] and therefore the transfers of OC may be decoupled from the transfers of water. Without data from marine end-members and a global model that incorporates the exchanges of rainwater, water vapor, dissolved OC and dissolved OC precursors together it is difficult to accurately assess how much OC is exchanged between the terrestrial and oceanic basins via precipitation. However, I argue that a large component of OC in marine rainwater is probably of marine origins and that net flux of OC from the continents to the ocean is smaller due to an appreciable ocean to land flux and oceanic studies that invoke an allochthonous input of terrestrial OC via rainfall must also consider an output.

Figure 2.

A conceptual diagram for the hydrologic transfers of water and organic carbon between the continents and ocean. Total water fluxes are in bold grey (103 km3 yr−1) located on the top of each arrow with the approximate land and ocean contributions given in parentheses, with the ocean flux coming second in italics. The water fluxes are adapted from Baumgartner and Reichel [1975], which provides a net flux of 40 for the net transfer of water from the ocean to land by precipitation that balances the river water flux. For this study, this net ocean-land flux was expanded to a gross flux of 64 by assuming that evapotranspiration is ∼60% of precipitation globally. The total fluxes of carbon are reported in bold black (in teragrams) under each arrow with the land and ocean contributions given in parentheses (the ocean flux come second in italics). These estimates assume that the fluxes of OC are dependent upon the amount of water being transported between and to each basin using compositions from this study and concentrations from this study and the literature. In order to obtain these carbon fluxes the concentration of OC associated with marine water moving over land was 25 μM, while the concentration of OC associated continental water moving over the oceans was 75 μM. The total flux of rainwater OC to the continents was made assuming a concentration of continental rain of 90 μM that is 75% continental OC (fossil fuel + modern terrestrial) and 25% marine. The total flux of rainwater OC to the ocean was made assuming a marine rain concentration of 32 μM that is 75% marine and 25% continental. The OC contributions in the recycled rain component for the continents and the oceans were therefore solved by difference. The approximate amount of fossil fuel OC deposited on land (explained in text) is provided inside the small factories. The riverine OC flux is for dissolved OC only.


[10] Thanks to excellent prep work by Kari Mull. Thanks also to Vivek Arora for insights into gross global precip budget and to J. Bauer and J. Cole for reading drafts of this ms. This work was funded by NSF EAR-0403962 and NSF DEB-0234504.