Greenhouse gas fluxes from Atacama Desert soils: a test of biogeochemical potential at the Earth’s arid extreme

Most terrestrial ecosystems support a similar suite of biogeochemical processes largely dependent on the availability of water and labile carbon (C). Here, we explored the biogeochemical potential of soils from Earth’s driest ecosystem, the Atacama

Please download to get full document.

View again

of 13
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 2 | Pages: 13

Extension: PDF | Download: 0

  Greenhouse gas fluxes from Atacama Desert soils: a testof biogeochemical potential at the Earth’s arid extreme Steven J. Hall  • Whendee L. Silver  • Ronald Amundson Received: 29 September 2010/Accepted: 3 September 2011   Springer Science+Business Media B.V. 2011 Abstract  Most terrestrial ecosystems support asimilar suite of biogeochemical processes largelydependent on the availability of water and labilecarbon (C). Here, we explored the biogeochemicalpotential of soils from Earth’s driest ecosystem, theAtacama Desert, characterized by extremely lowmoisture and organic C. We sampled surface soilhorizonsfromsitesrangingfromtheAtacama’shyper-arid core to less-arid locations at higher elevation thatsupported sparse vegetation.We performed laboratoryincubations and measured fluxes of the greenhousegases carbon dioxide (CO 2 ), nitrous oxide (N 2 O), andmethane (CH 4 ) as indices of potential biogeochemicalactivityacross thisgradient. Wewereable to stimulatetrace gas production at all sites, and treatmentresponses often suggested the influence of microbialprocesses. Sites with extant vegetation had higher Cconcentrations (0.13–0.68%) and produced more CO 2 under oxic than sub-oxic conditions, suggesting thepresence of aerobic microbial decomposers. In con-trast, abiotic CO 2  production appeared to predominatein the most arid and C-poor ( \ 0.08% C) sites withoutplants, with one notable exception. Soils were either aweak source or sink of CH 4  under oxic conditions,whereas anoxia stimulated CH 4  production acrossall sites. Several sites were rich in nitrate, and westimulated N 2 O fluxes in all soils by headspacemanipulation or dissolved organic matter addition.Peak N 2 O fluxes in the most C-poor soil (0.02% C)were very high, exceeding 3 ng nitrogen g - 1 h - 1 under anoxic conditions. These results provide evi-dence of resilience of at least some soil biogeochem-ical capacity to long-term water and C deprivation inthe world’s driest ecosystem. Atacama soils appearcapable of responding biogeochemically to moistureinputs, and could conceivably constitute a regionally-important source of N 2 O under altered rainfallregimes, analogous to other temperate deserts. Keywords  Arid    Atacama Desert    Chile   Global change    Methane    Nitrous oxide    Resilience   Wet-up Introduction A diverse suite of biogeochemical processes spans thebreadth of terrestrial ecosystems, apparently withremarkable consistency. Ecosystems ranging fromarctic peatlands to temperate grasslands to humidtropical forests all appear to support nutrient cyclesdriven by qualitatively similar interactions betweenmicrobial communities and ecosystem state factors,despite dramatic variation in process rates andfluxes. In the case of nitrogen (N), for example, S. J. Hall ( & )    W. L. Silver    R. AmundsonDepartment of Environmental Science, Policy,and Management, University of California at Berkeley,130 Mulford Hall #3114, Berkeley, CA 94720, USAe-mail:  1 3 BiogeochemistryDOI 10.1007/s10533-011-9650-7  mineralization of N from organic matter, bioticassimilation, nitrification, and denitrification all gen-erally occur across ecosystems, suggesting an appro-priate combination of abiotic conditions and a suitablebiotic community (Booth et al. 2005; Parton et al.2007).Eveninaridecosystemstypicallycharacterizedby aerobic biogeochemical processes, anaerobic pro-cesses such as denitrification (Peterjohn and Schle-singer 1991) or even methanogenesis (Peters andConrad 1995) can occur. Most ecosystems, however,arecharacterizedbyarelativeabundanceofseeminglycrucial biogeochemical resources: water and carbon(C). Desert ecosystems provide a principal exception,where rare pulses of water from rain, fog, or dewevents punctuate drought conditions and fuel brief periods of productivity and biogeochemical cyclingin fertile hot-spots and cryptogamic soil crusts(Austin et al. 2004). But even most deserts experiencemeasurable rainfall over an annual scale, and typicallysupport sparse communities of plants or cryptogams.What biogeochemical processes, in contrast, couldpersist in an ecosystem nearly devoid of water and C,with measurable rainfall only on scales of decades tocenturies, and lacking vascular plants or visiblecryptogamic crusts?Chile’s Atacama Desert contains such ecosystems,and provides a unique environment to test the resil-ience of soil biogeochemical processes to extremedeprivation of water and C. In the Atacama’s hyper-arid core, measurable moisture inputs rarely occur,vascular plants are entirely absent, organic C is scarce,and soil microbial communities would apparently beundetectableusinginstrumentationsent toMarsontheViking missions of the 1970s, leading some investi-gators to compare the Atacama with Mars (Navarro-Gonza´lez et al. 2003). Although some microbeshave recently been documented in these seeminglyinhospitable soils (Connon et al. 2007), their biogeo-chemical capacity remains uncertain. Fundamentalphysiological tradeoffs required for tolerance toextremely dry, nutrient-poor, or saline soils couldpotentially exclude the survival of microbial func-tional groups with particular metabolic traits, such asmethanogenesis or denitrification (Dubinsky 2008;Schimel et al. 2007). Yet, amelioratingabiotic stressescould promote biogeochemical processes if latentmicrobial capacity is present. Methanogenic archaeahave been cultured from desert soils after multipleweeks of incubation, even though they are normallyconsidered obligate anaerobic organisms (Peters andConrad 1995). It remains unclear, however, whethermicrobes can actually mediate biogeochemical pro-cesses in harsh terrestrial soil environments over thescale of days, given that water is likely available onlysporadicallyover annualtodecadaltimescalesatsomesites in the Atacama (McKay et al. 2003). Here, weasked if Atacama soils could respond biogeochemi-cally over timescales relevant to in situ water avail-ability,focusingontheproductionandconsumptionof greenhouse gases as characteristic indicators of C andN biogeochemical processes.The Atacama Desert spans a rainfall gradient fromarid (tens of mm rain year - 1 ) to hyperarid ( \ 2 mmrain year - 1 ) conditions, characterized by a switchfrom prominently biotic to prominently abiotic ele-mental cycling (Ewing et al. 2006, 2007, 2008). As such, the Atacama presents an ideal system for prob-ing the resilience of soil biogeochemical processes todrought and nutritional stress. Within the Atacama,microbial species richness and abundance declinesdramatically as rainfall decreases (Navarro-Gonza´lezet al. 2003; Maier et al. 2004; Drees et al. 2006; Lester et al. 2007). Navarro-Gonza´lez et al. (2003) reportedlow to negligible quantities of culturable microbes insoils in the hyper-aridregion, and postulated that thesesoils were too dry to support microbial life. Sub-sequent work, however, revealed the presence of sequenceable DNA in the sub-surface as well as viablebacterial colony-forming units (Connon et al. 2007;Lester et al. 2007; Maier et al. 2004). Stable iso- tope data suggests that rare precipitation events overmillennial time-scales stimulate pulses of organicmatter decomposition andgaseousNflux(Ewingetal.2007, 2008). To date, no experiments have examined soil biogeochemical capacity following water or Caddition. To address this, we sampled soils from arangeofsitesintheAtacamaalongaclimaticgradient,whichvariedwithrespecttoseveralkeyfactorsknownto influence trace gas production and consumption(rainfall, C, and N). Four sites supported no vascularplants, whereas three sites from the slope of the Andesco-occurred with sparse desert shrub and grasslandcommunities. In laboratory incubations, we addedwater, nutrient broth, and C substrates to soils, andvaried incubation headspace oxygen (O 2 ) concentra-tions to test soil biogeochemical potential. Ourresearch begins to address the hypothesis that biogeo-chemical capacity is essentially universal in terrestrial Biogeochemistry  1 3  soils, despite the myriad stressors they may experi-ence. We predicted that all sites would be able toproduce carbon dioxide (CO 2 ) and nitrous oxide(N 2 O), and produce or consume methane (CH 4 ) if supplied with favorable environmental conditions. Methods Site locationsWe collected soils that represent three unique geo-morphological units: hyperarid alluvial fans, salars(salt-encrusted lake beds), and an elevation transect of the western slope of the Andes (Table 1). Hyperaridalluvial fansoilsofMioceneandPliocene agefillmostbasins in the Atacama, and are characterized byextensive accumulations of sulfate (SO 4 ), nitrate(NO 3 ) and chloride, with minimal chemical weath-ering of the silicate matrix. We sampled a Miocene-aged soil located approximately 100 km from thecoast, and northeast of a location examined by manyprevious researchers (Navarro-Gonza´lez et al. 2003).In this region, rainfall events are extremely rare, andproduce measurable soil moisture on the scale of days.Fog and dew are more common, particularly withproximity to the coast, but do not contribute measur-ably to soil moisture (McKay et al. 2003). These soilshave a discontinuous desert pavement that overlies aporous 10-cm horizon consisting largely of SO 4 ,silicate dust, and other salts. Subtle micro-depressionsare characterized by a hard SO 4  crust with visiblecracks, which connect with subsurface pipes in the saltthat coalesce and emerge on downslope surfaces toform overland flow features. Soils sampled from thedepressions are termed ‘‘vernal pool’’ soils, given theevidence of historical water accumulation.Due to tectonic activity and/or hydrologic change,many basins are internally drained and contain‘‘salars,’’ consisting of dry lake beds encrusted withsalt. We collected samples from two salars. The Salarde Atacama is the largest salar in Chile, and receivesrunoff and salts from the adjacent volcanic peaksof the Andes. The sample selected here is from a SO 4 -rich zone. The Pan de Azucar is a small localdepression southeast of the Salar de Atacama fed byspring discharge. Notably, halite salts present in thesesoils are deliquescent. These salts absorb water vaporand produce liquid brine at a relative humidity of approximately 75%, potentially facilitating the sur-vival of cyanobacteria (Davila et al. 2008).The western slope of the Andes at * 20   S latitudeprovides a gradient of decreasing temperature andincreasingrainfallwithincreasingelevation,producingrecognizable vegetation zones. We sampled four loca-tions along this transect, from low to high elevation,loosely following the vegetation descriptions of Quadeet al. (2007): (1) pre-puna (essentially no vegetation),(2) puna (shrubs, succulent herbaceous plants, and C4grasses),(3)pre-altiplano(transitionbetweenshrubandC3 grass communities), and (4) altiplano (predomi-nantly C3 grass vegetation). Plant cover increasedsystematically along this gradient. This transect is veryclosetooneusedbyBarrosetal.(2008),whoexaminedsoil C, N and microbial activities with elevation.Soil sampling and analysisSoilsweresampledateachsitetoadepthof10 cmfroma single representative pit, yielding approximately2-kg samples that were subsequently homogenized. Table 1  Soil characteristics by site Soil (ID) Coordinates Altitude(m)Landform NO 3 -N( l g/g)NH 4 -N( l g/g)TotalN (%)TotalC (%)Org.C (%)C/N pH EC(mS cm - 1 )Vernal pool (VP) 24.020 S, 69.770 W 1001 Alluvial 2.7  9  10 4 160 2.4 0.02 0.02 0.01 7.6 412Salar de Atacama (SA) 23.659 S, 68.104 W 2301 Salar 1.7 0.2 0.01 0.03 0.03 3.0 8.0 406Pan de azucar (PA) 24.181 S, 68.777 W 2952 Salar 1.2  9  10 3 0.3 0.09 0.08 0.08 0.9 8.3 282Pre-puna (T1) 19.872 S, 69.407 W 2068 Andes 7.8 0.4 0.01 0.27 0.02 2.0 8.4 1.3Puna (T2) 19.883 S, 69.268 W 2734 Andes 4.2 0.5 0.03 0.17 0.13 4.3 8.3 0.2Pre-altiplano (T3) 19.898 S, 69.113 W 3374 Andes 4.3 1.3 0.05 0.46 0.46 9.2 6.4 0.2Altiplano (T4) 19.865 S, 68.807 W 4189 Andes 2.4 2.9 0.06 0.68 0.68 11.3 6.6 0.1EC denotes electrical conductivity Biogeochemistry  1 3  WemeasuredtotalsoilCandNbycombustiononaCEElantech CN analyzer (Lakewood, New Jersey), andmeasured organic C by difference after pre-treatingsamples with 0.1 M HCl to remove carbonates. Weextracted soluble NO 3  and NH 4  in a 1:5 slurry of soiland2 MKCl,followedbycolorimetricanalysisusingaLachat Quik Chem flow injection analyzer (LachatInstruments, Milwaukee, Wisconsin). We measuredsoil pH in a 1:2 slurry of soil and deionized water.Electricalconductivitywasmeasuredina1:1soil/waterslurry after filtration.Incubation treatmentsWe employed two treatments under ambient atmo-spheric conditions to determine the potential effectsof water or water plus C addition on greenhouse gasdynamics. For the first treatment we added deionizedsterilized water to achieve 65% of water holdingcapacity. For the second treatment we added steril-ized dissolved organic matter (tryptic soy brothsolution with 35% C, 9% N) at a concentrationsufficient to double ambient soil organic C at eachsite while raising soil moisture to 65% water holdingcapacity. We also explored the effects of redoxpotential on trace gas dynamics by incubating soilsunder sub-oxic and anaerobic conditions after addingwater as above.Replicate soil samples were incubated in the dark in 100 ml serum flasks. To minimize potentialmicrobial contamination, soils were transferred fromdouble-sealed ziploc bags to autoclaved serum flasksusing flame-sterilized utensils while positioned adja-cent to a flame, which created an updraft to reduce anyairborne contamination. Gravel fragments approxi-mately [ 2 mm were removed from soils with flame-sterilized tweezers to avoid any contamination fromsieving. Soils with strong structure (Pan de Azucar,Salar de Atacama) were broken up so individualaggregates were \ 2 mm. Flasks were capped withmicrobiological foam plugs between sampling periodsto allow gas exchange while preventing microbialcontamination. We incubated four replicate flaskscontaining 40 g of dry soil for each soil and treatment,except for the vernal pool soil, where only 20 g wasused because of sample scarcity. We determined thewater holding capacity of each soil by saturating 40 gof oven-dried soil with deionized water, allowingdrainage through filter paper, and measuring the massof water retained after 6 h (Miller et al. 2005).Prior to imposing treatments we equilibrated soilsin serum flasks for 2 days. After water or DOMaddition we capped vials with autoclaved butyl rubbersepta and measured gas fluxes over a 6-h period.Between flux measurements, we replaced septa withfoam plugs. Vials were incubated in the dark, and fluxmeasurements were repeated after three and 8 days.Gas concentrations were measured using a Shimadzu(Columbia, MD) gas chromatograph equipped with aflame ionization detector, thermal conductivity detec-tor, and electron capture detector (for measuring CH 4 ,CO 2 , and N 2 O, respectively). We also sampled fourempty vials at each time point to account for anyanalytical error or instrument drift. We do not reportCO 2  fluxes from DOM-amended soils because vialswith DOM alone showed a substantial CO 2  fluxresulting from thermal decomposition of autoclavedDOM; no N 2 O or CH 4  was produced or consumedfrom DOM alone, however.For the sub-oxic incubation, flasks were equili-brated under a N 2  atmosphere for 24 h prior to wateraddition, after which we added sterile, de-oxygenatedwater and measured fluxes as above. We repeated fluxmeasurements after 3 and 8 days. Oxygen concentra-tionsmeasuredwithanApogee(Logan,UT)O 2 sensorremained between 0 and 2% over the incubationperiod. For the anoxic incubation, we added water tosoilsandexposedthemtoanatmospherewith87%N 2 ,10%CO 2 ,and * 3%hydrogen(H 2 )inaglovebox.Thepresence of H 2  served to eliminate any O 2  by reactionwith a palladium catalyst, and may have served as anelectron donor to stimulate anaerobic metabolism.CO 2  fluxes were not quantifiable in the anaerobictreatment due to the presence of 10% CO 2  in theglovebox atmosphere. Gas concentrations were mea-sured after a 6-h incubation and compared withreplicate blank vials to calculate fluxes, and measure-ments were repeated after 3 days. Anaerobic incuba-tions lasted only 3 days to minimize any effects of H 2 on microbial growth.In a final experiment, we investigated the effects of waterandcarbohydrateadditionongasfluxesfromthevernal pool soil, the most arid soil that we examined.We imposed four treatments: water addition to 30, 65,and 100% of field capacity, and water addition to 65%of field capacity with dissolved glucose and sodiumacetate, for a total C addition of 2.5 mg C g - 1 soil. Biogeochemistry  1 3  Gas fluxes were measured under an ambient oxicatmosphere on days one, three, and eight as above. Forall experiments, we identified fluxes as statisticallysignificant when the mean of four replicates differedfrom zero according to a  t  -test ( a  =  0.05). Thesignificance of treatment effects on soil fluxes from agiven site were similarly evaluated with  t  -tests with a  =  0.05. Results The seven sites that we examined differed dramat-ically in soil chemistry (Table 1). The three Andeansites with extant vegetation (puna, pre-altiplano, andaltiplano) had higher soil C ( [ 0.1% C) than the foursites where vegetation was absent (vernal pool,salars, and pre-puna  \ 0.1% C). Conversely, thevernal pool and pan de azucar sites had extremelyhigh inorganic N concentrations, on the order of 1–27 mg N g - 1 soil, yielding soil C/N ratios \ 1. Thevernal pool, pan de azucar, and salar de Atacamasites were relatively saline (electrical conductiv-ity [ 200 mS cm - 1 ), reflecting high concentrationsof NO 3  and SO 4  salts.Soil CO 2  fluxes varied according to the presence of extant vegetation. Fluxes from the three vegetatedsites exceeded those from the non-vegetated sites by2–3 orders of magnitude, and declined over time inboth incubations (Fig. 1a). Most sites showed netpositive CO 2  fluxes after wet-up under oxic condi-tions, except for the two salars, while all sites hadpositive CO 2  fluxes under sub-oxic conditions. Thesub-oxic treatment significantly decreased CO 2  fluxesfrom the relatively C-rich sites (puna, pre-altiplano,and altiplano) to 71, 26, and 32% of the oxic treatmentflux values compared on the first day of sampling(  p \ 0.001, pairwise  t  -tests). Likewise, CO 2  fluxesfrom the hyper-arid vernal pool site decreased signif-icantly (  p \ 0.01) under a sub-oxic headspace to 67%of the oxic CO 2  flux on the first sampling. For the restof the sites, CO 2  fluxes remained similar amongheadspace treatments, or subtly increased under thesub-oxic treatment (Fig. 1b). The salar sites, however,exhibited significant negative CO 2  fluxes after initialwet-up under oxic conditions ( - 0.013  ±  0.001 and - 0.010  ±  0.003  l g C g - 1 h - 1 , respectively), poten-tially due to CO 2  consumption from carbonate forma-tion. In contrast, these two sites showed small netpositive CO 2  fluxes under sub-oxic conditions wherebackground CO 2  concentrations were low ( \ 40 ppm).Fluxes of N 2 O varied significantly among sites andtreatments, but all soils showed the capacity for N 2 Oproduction (Fig. 2, Appendix). Under the oxic treat-ment, the three vegetated sites produced N 2 O fluxesthat exceeded the other sites by 1–2 orders of magnitude (Fig. 2a), while the four C-poor sitesproduced very small N 2 O fluxes ( \ 0.01 ng g - 1 h - 1 )that often did not statistically differ from zero(Appendix). Adding DOM stimulated N 2 O productionin several soils, most dramatically in the non-vege-tated and extremely C-poor pre-puna soil, where N 2 Ofluxes increased by three orders of magnitude on the VPSAPAT1T2T3T4    C   O    2   -   C   f   l  u  x   (  µ  g  g   -   1    h  r   -   1    )    0 .   0   0 .   5   1 .   0   1 .   5 Day 1Day 3Day 8 VPSAPAT1T2T3T4 Soil/Date    C   O    2   -   C   f   l  u  x   (  µ  g  g   -   1     h  r   -   1    )    0 .   0   0 .   5   1 .   0   1 .   5 Day 1Day 3Day 8 ab Fig. 1  Mean CO 2 -C fluxes ( ± SE) by soil, sampling date, andthe oxic ( a ) and sub-oxic ( b ) incubations, respectively.  Shaded bars  represent different sampling dates for each soil, and soilabbreviations are described in Table 1. Soils to the  right   of the dashed line  through the  x -axis support vascular plantsBiogeochemistry  1 3
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks