Reviews

Global change Soil microbiomes and

Janet K. Jansson * and Kirsten S. Hofmockel * Abstract | The soil microbiome governs biogeochemical cycling of macronutrients, micronutrients and other elements vital for the growth of plants and animal life. Understanding and predicting the impact of climate change on soil microbiomes and the services they provide present a grand challenge and major opportunity as we direct our research efforts towards one of the most pressing problems facing our planet. In this Review, we explore the current state of knowledge about the impacts of climate change on soil microorganisms in different climate-sensitive soil , as well as potential ways that soil microorganisms can be harnessed to help mitigate the negative consequences of climate change.

Permafrost Over the past century, CO2 levels have steadily increased emissions, thus exacerbating warming trends. This Soil that has been frozen for at and global temperatures have risen accordingly. As out- is a concern because the total amount of , least 2 consecutive years. lined in the recent US national climate assessment1, the including within , is estimated to be ~3,300 climate is predicted to continue to change with weather petagrams (Pg) — approximately five times larger than Carbon use efficiency (refs8,9) The difference between the patterns becoming more erratic and extreme. As soil the current atmospheric pool of CO2 . However, amount of carbon respired as microorganisms are largely responsible for cycling of there is considerable uncertainty in climate models on

CO2 and that incorporated into soil organic carbon (SOC) and other nutrients, they whether this pool of soil carbon will increase or decrease the cellular biomass. have a key role in climate feedback, including produc- in the future. The majority of empirical data from cli-

tion or consumption of greenhouse gases such as CO2, mate change studies in the field have relied on measure-

CH4 and N2O. However, whether soil will become a ment of changes in . In order to improve source or sink of greenhouse gases under future climate models of soil carbon–climate feedback, there is also a scenarios has been difficult to predict2,3 due to unknown need to determine how bulk soil carbon stocks change changes in soil carbon and nitrogen pools, and differ- with changes in climate10. ences in microbial responses between soil locations. Soil microorganisms carry out the dichotomous roles Thus, although the importance of soil microbial ecol- of mineralization of SOC and stabilization of carbon ogy for prediction of future climate impacts has been inputs into organic forms. The balance between these

recognized, it remains a challenge to integrate with two processes governs the net flux of CO2 and CH4 to landscape-scale climate models4. the atmosphere. The proportion of carbon substrate In this Review, we provide an overview of research that is retained in the microbial biomass compared

describing responses of soil microorganisms to the fol- with that respired as CO2 is referred to as the microbial lowing anticipated changes in climate1: elevated levels of carbon use efficiency. Heterotrophic respiration of SOC

atmospheric carbon dioxide (eCO2); elevated tempera- has globally increased as a result of climate change, thus

ture; increased ; increased precipitation and/or contributing to increased atmospheric inputs of CO2 flooding; and increased fire frequency. Although dis- (ref.11). However, losses of soil carbon to the atmosphere cussed independently, we also emphasize where com- could be countered by increased soil carbon inputs due pounding disturbances occur due to a combination of to increased plant growth12 and autotrophic fixation by climate change impacts5; for example, heat waves can soil microorganisms. Also, the temperature sensitivity 6 include both eCO2 and increases in temperature , and of decomposition of SOC varies with the quantity and there can be combined effects of drought and warming7. chemistry of plant litter and pre-existing SOC13. Thus, This knowledge should elevate our understanding of even within specific biomes, the local biogeochemical Earth and Biological Sciences how microbial communities and the ecosystem services environment strongly influences microbial metabolic Directorate, Pacific Northwest they provide are influenced by the combined pressures responses to climate (Box 1). Developing a mathematical National Laboratory, of changing climate and compounding disturbances, understanding of the microbial ecology that drives eco- Richland, WA, USA. which in turn will be useful for protecting, managing system carbon use efficiency and the feedback with cli- *e-mail: [email protected]; and mitigating ecosystem resilience. mate forcing is therefore a pressing need for improving [email protected] One major concern with climate change is that soil climate change models. https://doi.org/10.1038/s41579- microorganisms will mineralize more SOC and con- Nitrogen cycling also has a strong influence on cli- 019-0265-7 tribute substantially to (CO2 and CH4) mate change, because nitrogen availability is closely

naTure RevIeWs | MiCrobioloGy volume 18 | January 2020 | 35 Reviews

correlated to the production of CO2, N2O and CH4 Current status and knowledge gaps (ref.14) . In particular, management of N2O emissions Soil represents one of the most highly diverse ecosystems

remains a pressing issue because N2O is a potent green- on our planet with an interacting community of bacteria, house gas, with 298 times the warming potential of archaea, viruses, fungi and protozoa: collectively referred

CO2, and emissions from soil represent 56–70% of all to as the ‘soil microbiome’. The soil abiotic environment 15 global N2O sources . Managing N2O emissions within is also highly heterogeneous, with disconnected air-filled soil ecosystems has been a formidable challenge due to and/or water-filled pores, and patchy resources that can the complexity of the microbial communities involved16 serve as hotspots for microbial growth21. When com- and the spatial and temporal variation of the soil envi- bined with the influence of plants and soil fauna (for ronment17,18. Recently, microorganisms were discov- example, insects and earthworms) as well as changes in ered that can completely oxidize ammonia to nitrate soil moisture, temperature and fluctuating redox states, 19 (comammox) without generating N2O as a by-product, the soil environment is highly dynamic. However, cli- and these are abundant in soil20. This finding illustrates mate change is introducing more extended and higher the importance of gaining a better understanding of soil extremes of change with unknown consequences on microbial physiology and metabolism and how those the stability and resilience of the soil microbiome22. processes are impacted by climate change. Therefore, better understanding of microbial traits that

Box 1 | Climate-sensitive soil ecosystems The The Arctic is one of the most climate-sensitive regions on Earth, with average temperatures increasing at nearly twice the global rate138, resulting in drastic changes in the landscape including permafrost thaw, changes in precipitation patterns and changes in vegetation. With permafrost thaw, microorganisms become more active and start to decompose the enormous reservoir of stored carbon (~1,300–1,580 Pg of carbon)139, which is approximately half of the total soil carbon stock. Recent estimates suggest that ~5–15% of the carbon currently contained in permafrost is susceptible to microbial 139 decomposition, resulting in a substantial source of CO2 emissions over the next few decades . Although this is less CO2

than that emitted from current fossil-fuel emissions, adding CH4 to the equation greatly accelerates the warming impact 140 because CH4 has 34 times the climate-forcing influence than CO2 (ref. ). As the huge reservoir of carbon in high latitude soils is released, positive feedback to climate change is expected at the planetary scale. Forests Forests, ranging from boreal to temperate to tropical, together cover ~30% of the total land surface and have important roles as soil carbon sinks with large amounts of stable organic matter141. It is possible that predicted losses of soil carbon to the atmosphere due to increased microbial decomposition of soil organic carbon under warming will be partly countered by increased carbon inputs due to increased plant growth142. However, with increasing temperature, drought

severity and fire frequency, forest ecosystems have the potential to transform from net sinks to net sources of CO2 in the future142. Both fungal and bacterial communities in forest soils have been shown to respond to changes in climate, but the types of microorganisms and their specific responses differ among forest ecosystems. This is partly because of the differences in the litter type and quality depending on the plant communities (for example, coniferous versus deciduous forests) and differences in soil pH142. Grasslands Grasslands comprise ~26% of the global land area and store an estimated 20% of the total soil carbon stock143. The grassland soil carbon pool is large, approximately one or two orders of magnitude larger than the above-ground plant biomass pool, due to the deep and abundant rhizosphere that deposits carbon in the soil144. As a result, the projected vulnerability of grasslands to changing climate is intimately related to plant–microorganism interactions occurring in the rhizosphere, and to bulk soil processes that cycle carbon and other nutrients. With climate change, most grassland ecosystems are experiencing increased periods of drought and fire, mixed with more periodic and extreme precipitation events. The resulting changes in soil moisture influence both the above-ground plant growth and the microbial community compositions and functions in soil. Owing to differences between soil types and plant cover across grassland ecosystems, it has proven difficult to generalize what the long-term impacts will be on microbial community functions and climate feedback. Wetlands Wetlands interface terrestrial and aquatic systems, resulting in various ecosystems amenable to microbial greenhouse

gas production. Wetland CH4 emissions are the largest natural source of CH4 and contribute approximately one-third of total emissions145. Presently, conventional greenhouse gas mitigation policies do not include feedback associated with 146 wetland CH4 emissions and could be improved with more accurate representation of microbial responses. Water

availability is generally a strong predictor of CH4 emissions: reduced precipitation enhances O2 availability, promotes

organic matter decomposition, increases CO2 release and reduces CH4 emissions, whereas increased precipitation and

more anaerobic conditions favour CH4 production. Drylands Deserts and other dryland soils are characterized by water deficiency that restricts plant and microbial activity. Owing to the expanse of global arid regions (approximately one-third of the planet’s surface), they collectively store ~27% of the total terrestrial organic carbon stocks147. Climate change is resulting in an expanse of soil desertification and dryland areas are projected to increase by 11–23% by the end of this century76. These changes in water availability can have profound and lasting impacts on soil microbiomes. However, because arid soil ecosystems are globally distributed, it is difficult to generalize microbial responses to increasing drought across soils and regions.

36 | January 2020 | volume 18 www.nature.com/nrmicro Reviews

microbiome, such as the abundance of fast-growing,

CO2 opportunistic ‘r-strategists’ compared with slow-growing ‘K-strategists’, as well as environmental properties can help to predict how the soil microbiome will respond to Current state Desired outcome different climate change scenarios26. High-throughput sequencing has been instrumental in revealing the microbial diversity and composition in various soil ecosystems27, providing a valuable baseline for comparison as the climate changes. It is now recog- nized, however, that compositional information does not always inform function (Fig. 1). Not all members of a community, or even cells within a specific popu- lation, are active at any given time. Activity is governed by a complex interaction of gene regulation that gov- • Sequence the soil microbiome to • Understand the molecular details of erns which genes are expressed and access to resources. identify what is there and what their soil biochemical reactions that are Variability in moisture, temperature and local atmos- potential functions are responsible for cycling carbon and pheric chemistry within the soil impacts the phenotypic • Challenging to determine function other nutrients beyond measurement of bulk-scale • Make better predictions of climate response of the soil microbiome with feedback to cli- processes, such as soil respiration change impacts on metabolic mate change. The diverse genetic potential within the • Genes, proteins and metabolites pathways involved in carbon and nutrient • Determine the metaphenome of the soil microbiome interacts with environmental shifts to metabolic pathways are largely soil microbiome under different induce microbial gene expression. This collective pheno- unknown climate conditions typic output of the microbiome, the metaphenome, gener- ates ecosystem-scale elemental cycling28. Understanding the parameters connecting local microbial phenotypes Fig. 1 | Current state and desired outcome of soil microbiome science. Owing to to larger-scale ecosystem responses is thus an important advances in sequencing technologies, it is currently possible to determine the taxonomic composition of soil microbial communities and to determine how climate change frontier for improving climate models and for managing influences the community membership. Understanding the details of biochemical soil microbiomes in response to climate change (Fig. 1). pathways, such as those involved in soil respiration, that are carried out by interacting Currently, little is known about the fundamental members of soil communities and how key functions are impacted by changes in climate microbial-scale mechanisms that control ecosystem-scale are crucial science frontiers that can be addressed with multi-omics approaches and responses to climate change. Soil microorganisms do not other emerging technologies. Colours in the left panel represent different microbial cells respond to mean environmental conditions, but rather and in the right panel represent different steps in biochemical pathways. to instantaneous conditions at the microscale that trigger biochemical reactions, microbial responses and meta- confer ecosystem resilience to climate change is needed bolic interactions. Biogeochemical responses to abrupt for predicting and managing ecosystem responses to environmental change often include temporal lags as climate change. the soil microorganisms adapt. By contrast, a gradual change, such as an increase in temperature, allows more Influence of the soil microbiome on emergent ecosystem time for evolution by selection for species or genes that properties. As soil habitats are dynamic systems, most enable resistance to heat and associated stress conditions. soil microorganisms have evolved strategies to cope with The historical context also has a role in the community changing environmental conditions. Generally, as envi- response. One study found that the historical exposure ronmental conditions shift, the resident microorganisms of grassland and adjacent forest soils to changes in soil either adapt, become dormant or die. Soil microorgan- water content influenced the ability of the soil micro­ isms respond to environmental stress in different ways, biome to respond to new changes in soil moisture29. depending on their genetic and physiological states23. These findings imply that different soil ecosystems are Acclimation and adaption to change is dependent sensitive to climate change to different degrees. r-Strategists on the degree of perturbation and the time necessary Species that typically have high growth rates and are able to to regulate gene transcription and translation and/or Influence of the soil environment on microbial responses respond quickly to resources accumulate mutations or new genes through horizontal to climate change. As soils vary tremendously with as they become available. gene transfer. However, at present, quantifying microbial respect to their biotic and abiotic properties, it is difficult physiological responses (for example, drought resistance, to generalize the impact of climate change on soil micro- K-Strategists dormancy or reactivation) remains a major gap in biomes across different soil ecosystems (Box 1). Within Species that typically are slow 24 growing and adapted to utilize modelling ecosystem responses to change . a specific soil class, there are differences in biogeo­ minimal resources. Changes in microbial community structure may chemistry that govern the types of microorganisms that influence the stability and resilience of the commu- are present, including pH30 and salinity27. In addition, the Metaphenome nity to future disturbances. Climate change will impact soil structure and the soil moisture content influence A community phenotype that 31 is the product of genomic interactions between microbial populations in com- the creation of microbial habitats and niches with cas- potential encoded in munities which in turn impact the ability of single spe- cading effects on carbon and nutrient transformations. metagenomes and the cies to adjust25. There can be mismatches between the Therefore, understanding the fine-scale distribution and environmental conditions that responses of different species to rising temperature, for connectivity of soil microbial communities is required govern which genes are example, that can lead to changes in their dispersal pat- to better understand how climate change influences expressed. terns. Focusing on specific functional traits in the soil species interactions and metabolism32. For example, this

naTure RevIeWs | MiCrobioloGy volume 18 | January 2020 | 37 Reviews

C4 plants knowledge is important for carbon cycling because the differently to eCO2, influencing the amount and types

Plants that fix CO2 into a way that microbial species allocate carbon will ultimately of carbon inputs to the rhizosphere. For example, four-carbon compound (in determine whether carbon persists in soil31 and how C4 plants are more efficient at photosynthesis than addition to a three-carbon changes in environmental conditions influence these C3 plants41 and may allocate more carbon below ground compound) and that have high photosynthetic efficiency due processes. Although it is well established that consortia of to rhizosphere-associated microorganisms, resulting 34 to an absence of microorganisms within various soil niches interact and in shifts in community composition . The increased photorespiration. respire CO2, N2O and CH4, the energetics and thermo- rhizodeposition in response to eCO2 can ‘prime’ micro- dynamics of the organic carbon electron acceptors that bial decomposition of existing SOC42. Priming is the C3 plants drive microbial metabolism are not well characterized in stimulation of decomposition of old SOC through Plants that fix CO2 into a three-carbon compound and the context of the soil habitat. Our present frontier is to the addition of new microbial substrates that could that have a lower photosynthetic describe the physiological response surface (that is, the include root exudates and/or increased litter inputs, efficiency than C4 plants. metaphenome) of soil microbial communities among both of which may be enhanced by eCO2. A study com-

the soils of our planet. bining meta-analysis and modelling revealed that eCO2 initially stimulates photosynthesis and carbon inputs to

Impacts of climate change soil. However, over decadal timescales, eCO2 increased In this section, we consider different types of community the microbial decomposition of SOM43,44. Predicting the and physiological responses that soil microorganisms use balance between carbon accrual through mineral associa­ to cope with changing environmental conditions caused tion and soil aggregation45 and accelerated decomposi- by climate change. Although it is not possible to general- tion through priming46 remains an additional challenge. ize across different terrestrial ecosystems, due to different This is because changes in soil carbon stocks are diffi- predicted climate change variables across geographical cult to detect47 and the fundamental biology regulating regions (Box 1), we provide some examples for context. SOM decomposition has not been discovered. A recent comparison of priming effects in temperate and tropical

Elevated CO2. Several eCO2 field experiments have pro- forest soils found that the amount of soil carbon released vided valuable data about microbial responses to this by priming was not proportional to the rates of soil res-

anticipated change in climate. The free-air CO2 enrich- piration, due to differences in SOC turnover rates in the ment experiments were set up across a range of ecosys- different soil biomes39. These results suggest that prim- tem types to compare long-term exposure to elevated ing is influenced both by the amount of organic matter

and ambient CO2 levels. Several studies have demon- deposition and the rate of SOC turnover.

strated shifts in the microbiome with eCO2. A 10-year It is difficult to uncouple the compounding effects of

cross-biome study found ecosystem-specific responses elevated eCO2 from warming, as can also be seen in our as well as common responses of soil bacteria, such as discussion of temperature impacts below. For example, (ref.33) increases in Acidobacteria with eCO2 . In Australian with an increase in soil moisture due to eCO2, there can 34 grasslands, eCO2 resulted in shifts in archaea and fungi, also be an increase in warming, which dries the soil . In and specific groups of bacteria34. Efforts are being made the Australian grassland study34, although total fungal

to understand how phylogenetic shifts represent eco- abundance increased under eCO2, when eCO2 was in logical traits of microbial populations35. A trait-based combination with warming there was a decrease. The

approach provides a framework for integration of long-term effects of eCO2 on soil carbon stocks also microbial physiology into ecosystem ecology. depend on water and nutrient availability, which influ- Changes have also been found in potential func- ence photosynthesis, microbial decomposition and the

tions carried out by the soil microbiome under eCO2 net accrual of soil carbon. Understanding how changes

through screening gene abundances in metagenomes. in CO2 interact with other important environmen- For example, the BioCON grassland experiment revealed tal variables including temperature, precipitation and

eCO2-stimulated increases in gene families associated nutrients (such as phosphorus) is therefore crucial for with decomposition, nitrogen fixation and dissimilatory predicting microbiome responses among soil ecosys- nitrate reduction, and lower abundances of gene families tems. This highlights the need for continued and future involved in glutamine synthesis and anaerobic ammo- long-term field studies that apply the suite of major 36 nium oxidation . In arid grasslands exposed to eCO2, climate-associated changes that are likely to occur in a microbial genes involved in decomposition, nitrogen fixa- given region.

tion, carbon fixation, CH4 metabolism, nitrogen minerali- zation and denitrification were all increased37. Knowledge Increased temperature. Temperature determines the of shifts in gene functions involved in soil organic matter growth rates and yields of pure cultures of microorgan- (SOM) cycling enables a deeper understanding of how isms. Some physiological responses of microorganisms

microorganisms are affected by eCO2. However, because to higher temperatures include changes in lipid com-

eCO2 field experiments have not been performed globally, positions of cell membranes to reduce membrane flu- with replica data sets, it remains challenging to inform idity and the expression of heat shock proteins. It has global terrestrial ecosystem models22. been more difficult to assess the temperature response Plant–microorganism interactions are an impor- of soil microorganisms in situ, although advances in tant scientific frontier for quantifying carbon exchange sequencing and functional gene arrays have revealed

between the atmosphere and the soil. eCO2 can enhance community and functional gene shifts in response to plant biomass, allocation of carbon to roots and soil increased temperatures in the field48–50. The community microbial activities38–40. Different plant species respond and physiological responses of the soil microbiome to

38 | January 2020 | volume 18 www.nature.com/nrmicro Reviews

higher temperatures also depend on the biome under three main responses to 12 years of warming in grassland study (for example, forest compared with grassland). For soil experimental plots as follows: changes in microbial instance, warming has been shown to have contrasting community structure, largely driven by changes in the impacts on soil fungi in different boreal forest ecosys- plant community structure (there was a shift to favour tems, resulting in either stimulation51 or suppression52 growth of C4 plants over C3 plants); differential impacts of fungal biomass and activity; these differences are on bacteria, but not fungi; and enhanced nutrient cycling presumably due to differences in soil moisture and/or that can feedback to promote plant growth59. vegetation at different sites52. The resilience of microbial communities to higher To determine the impacts of extended soil warming temperatures is ultimately dependent on the compound- on the temperate forest soil microbiome, the Harvard ing impacts of drought, warming and plant type. The

Forest Ecological Research Station Long Term Ecological Prairie Heating and CO2 Enrichment (PHACE) exper- Research site carried out a long-term soil warming iment, on Wyoming grasslands, studied the impacts of 37 experiment where the soil was warmed by 5 °C above 12 years of eCO2 in combination with warming . Genes ambient temperature for up to 26 years48. One of the involved in carbon and nitrogen cycling were enriched

major findings of this long-term study was the acclima- under eCO2 alone and in combination with warming. tion of microbial respiration and associated microbial However, nitrogen cycling was suppressed under warm- mechanisms in four phases: rapid carbon loss through ing alone. The positive plant community response that respiration; microbial community reorganization; a shift resulted in increased biomass compounded the effect towards a more diverse, oligotrophic microbial commu- of shifts in precipitation7,55. Therefore, although warm- nity with higher soil respiration in heated plots than in ing stimulated both the carbon input into soils and soil controls; and a reduction in more recalcitrant carbon respiration, carbon loss through respiration was mainly pools with an anticipated further change in microbial offset by the increased plant biomass. Together, these community structure48. Over the short term, the appar- responses would act to weaken the positive feedback ent acclimation of soil respiration was attributed to to climate warming and reduce soil carbon loss. In reduced microbial biomass and thermal adaptation of summary, although most climate models predict pos- soil respiration10. Yet the physiological adjustments of itive feedback as a result of warming due to increased individual populations remain to be quantified in the soil respiration and a decrease in soil storage2,60, there field context. New isotopic methods provide opportuni- are confounding experimental results that are largely ties to bridge this knowledge gap and quantify microbial ecosystem dependent. population dynamics under field conditions53. The Harvard Forest warming experiment resulted in Permafrost thaw. A serious consequence of global an initial loss of soil labile carbon, followed by increased warming is the thaw of permafrost soils in the Arctic. degradation of more recalcitrant carbon compounds48. As permafrost soils store a huge reservoir of carbon The authors estimated a loss of ~710 g of carbon per (Box 1), the potential feedback to the climate upon per- square metre of soil and, by extrapolation, a loss of mafrost thaw is also huge61. Climate change is resulting 190 Pg of carbon by the end of the century with con- in increases in the depth of the seasonally thawed active tinued warming trends, which is comparable with layer at the expense of the underlying permafrost. As the amount produced over the past two decades from the permafrost thaws, liquid water becomes more avail- fossil-fuel emissions48. Therefore, sustained warming able and microbial activity increases. This can result in for 26 years resulted in depletion of SOC with corre- an increased decomposition of SOC and an increase in (ref.62) sponding reductions in microbial biomass, suggesting production of the greenhouse gases CO2 and CH4 . deleterious consequences of long-term warming for A general feature of permafrost thaw is a change in soil sustainability48. The reduced carbon availability soil moisture that largely governs microbial activity. For

corresponded with decreased abundance of fungi and example, the ability to produce CH4 is a distinguishing Actinobacteria, and an increased abundance of oligo­ feature of certain thawed permafrost environments63–65, trophic bacteria54, reinforcing the idea that microbial depending on the landscape hydrology64, soil depth and traits may be associated with ecosystem shifts in respira- redox conditions66. Experimental warming in the field tion. These studies provide further evidence that under- resulted in a lower redox potential at the permafrost standing the physiology of microbial populations and boundary and an increase in methanogens as a result66. communities can enhance our predictive understanding By contrast, the redox conditions in mineral permafrost of ecosystem responses to climate change. layers can favour iron reduction. A recent landscape tran- In contrast to the Harvard Forest study, experimen- sect of discontinuous permafrost in Alaska suggested the tal warming of different grassland soils for 2 years7 or importance of Fe(II) content as a direct driver of micro- 3 years55 resulted in an increase in microbial biomass. bial community composition67. This hypothesis is sup- However, both forest and grassland studies found ported by the finding of abundant proteins corresponding shifts in microbial community compositions, includ- to iron-reducing bacteria in Alaskan permafrost63. ing decreases in fungal-to-bacterial ratios48,55–57 and Several recent studies have used molecular approaches increases in Gram-positive bacteria34,48,58. The decrease to unravel microbial community responses to perma- in fungi with warming could reflect the fact that bacte- frost thaw63–65,68. Metagenome sequencing revealed that ria have traits that provide a competitive advantage at the microbial community membership and functional higher temperatures, including faster growth rates and potential in permafrost are dissimilar from those in the better nutrient competition. One study summarized active layer63,64,68 and that the permafrost microbiome

naTure RevIeWs | MiCrobioloGy volume 18 | January 2020 | 39 Reviews

changes rapidly upon thaw68. In addition, permafrost result in declines in microbial functions that are impor- microbiomes are not the same everywhere and different tant for ecosystem sustainability77. As soil becomes drier, species compositions have been found across sampling there is less water in soil pores, resulting in disconnected locations in the Arctic64,69. Most studies find an increase resource islands; subsequently, less SOC is decomposed (ref.78) in Actinobacteria with depth into the permafrost, but the and respired to CO2 . Together, these factors inter- species can differ depending on location64,69. To under- act to generate responses that range from decreased pro- stand how permafrost microorganisms are genetically ductivity under drier conditions79 to reduced carbon loss equipped to survive in permafrost and respond to thaw, through suppressed respiration60. metagenome-assembled genomes (MAGs) have been ass­ As arid soil ecosystems are widely distributed, rang- embled from permafrost metagenomes, including from ing from hot and cold deserts with limited annual rainfall previously uncharacterized taxa64,65. MAGs from sea- to Mediterranean grasslands with dry summers and wet sonally thawed active layer soils revealed traits that are winters, it is difficult to generalize microbial responses needed for survival of discontinuous freezing and thaw- to increasing drought with climate change. For exam- ing cycles, including the capacity for production of cold ple, life in desert soils is often constrained due to carbon and heat shock proteins, cryoprotectants and DNA repair and moisture limitation, and is therefore dominated by mechanisms64. Examination of the MAGs also provided surface-dwelling photoautotrophs80. This results in the clues as to how specific members of the permafrost soil formation of biological soil crusts (biocrusts); for exam- microbiome respond to changing resource availability ple, >40% of drylands have biocrusts81. Biocrust popu- with permafrost thaw. For example, MAGs obtained from lations (for instance, cyanobacteria and lichens) carry a permafrost thaw gradient had genes required for degra- out fixation of carbon and nitrogen, and are often the dation of plant polysaccharides, including cellulases and major primary producers where these elements form. xylanases65, suggesting that they were poised to degrade The concern is whether biocrusts will be able to adapt to these plant-derived substrates as they became available. In more intermittent rainfall and harsher and more extreme another study, MAGs containing genes for carbohydrate conditions in the future82,83. There are also major reper- metabolism increased after 4.5 years of 1 °C of warming cussions for soil degradation because biocrusts bind and in the field66. Several MAGs correspond to methanogens, stabilize the surface soil. in particular, in wetter locations where methanogenesis In grassland ecosystems, drought can have long- 64–66 occurs . However, a diverse capacity for CH4 oxida- lasting impacts on the soil microbiome because of shifts tion has also been revealed in MAGs, suggesting genetic in vegetation to more drought-tolerant plant species and

mechanisms for consumption of CH4 from thawing their subsequent selection for different root-associated permafrost before it reaches the atmosphere70. microorganisms26,85. Network analyses in mesocosms84 These molecular studies are helping to piece together and multi-year field experiments85 revealed that bacteria genomic-level details of inter-kingdom responses to are more sensitive to drought than fungi in grasslands. changing conditions in the Arctic. Interestingly, thou- Fungi thus potentially contribute to the maintenance sands of viral sequences were found in Arctic meta­ of carbon and nitrogen cycling when water is scarce86. genomes and some of them included auxiliary metabolic Also, as soils become drier and microbial dispersion genes for metabolism of plant polymers and numer- becomes more constrained within physically protected ous host linkages to diverse carbon cycling micro­ soil pores87,88, fungal hyphae may help to bridge spatially organisms71. Viral sequences were shown to differ across discrete resources89, which may help the bulk micro­ a permafrost thaw gradient, with a shift from ‘soil-like’ biome as well. Thus, it is important to better understand viruses in the drier soils to ‘aquatic-like’ viruses in the how inter-kingdom interactions generate community wetter soils72. Several of the viruses could be linked to responses to drought stress. potential bacterial hosts, many of which are key players Soil microorganisms have evolved various physio- in SOM decomposition72. Together, these findings sug- logical strategies to cope with drought stress, such as gest that both bacteria and viruses may have a role in osmoregulation, dormancy or reactivation and extra- carbon turnover in thawing permafrost. Although less cellular enzyme synthesis90. In order to survive under Metagenome-assembled is reported about fungi in permafrost, there have been lower water matric potentials, microorganisms accumu- genomes reports of an increase in some taxa in sequence data, late solutes (osmolytes) to retain cell turgor78. However, (MAGs). Genomes that are including mycorrhizal fungi following thaw73. One osmolyte accumulation might be too energetically derived from assembled study found differences in the fungal community com- expensive under intense drying conditions91,92. Soil metagenome data; often using a process called ‘binning’. positions in the rooting zone of permafrost compared microorganisms might simply persist in a dehydrated with an adjacent water-saturated bog, with an increase state and recover and regrow when moisture becomes Auxiliary metabolic genes in putative saprotrophic and pathogenic fungi follow- available93. Another physiological strategy is the pro- Genes on viral sequences ing thaw74. Functionally, these shifts in fungal sequences duction of extracellular polymeric substances to retain (genomes or contigs) that 29 represent non-viral metabolic were mainly correlated with associated shifts in plant water at low matric potentials . Also, members of some 74 genes, such as genes involved types due to warming . bacterial taxa, such as Actinobacteria and Bacilli, can in carbon metabolism. persist in drought-impacted soils94,95 because of their Drought. Drought is expected to be the major conse- ability to conserve activity and become dormant under Matric potentials quence of future climate change in mesic grassland eco- dry conditions. The potential energy of water 75 (Box 1) that is due to adhesion of systems . In addition, increased desertification Another physiological hurdle for arid soil micro- 78 water molecules to soil is predicted for most semi-arid or arid regions in the organisms is how to cope when the soils are re-wet . particles. coming decades76. Increasing drought is predicted to As microbial activity diminishes, extracellular enzyme

40 | January 2020 | volume 18 www.nature.com/nrmicro Reviews

activity may persist, causing bioavailable substrates to wetting, suggesting that a sizeable portion of the wet-up accumulate, facilitating reactivation when moisture response could be driven by degradation of dead micro- returns90,96. Re-wetting a dry soil results in a pulse of bial cells98 and hinting at a role for viral predation of bac- microbial activity — the so-called Birch effect97. High teria in this response. One study compared modelling mortality has been observed for bacteria and fungi upon simulations and empirical data to determine that soil microbial diversity was higher under dry conditions99, presumably because there were more disconnected soil niches in the dry soil (Fig. 2). Following wetting, there was Incr enhanced connectivity, more dispersal, more anaerobic O 2 eas d C ed ate tem niches and a sudden increase in nutrients, resulting in lev pe CO2 E ↑SOC ra t CO2 an increase in anaerobic taxa and a decrease in diversity. ur e The community was, however, resilient and returned to CH4 CO 99 CO 2 its previous state following re-drying . By contrast, in 2 ↓SOC a California grassland, the microbial biomass increased

P e during the dry season, and then declined during the wet r m 100 a season . A comparison of different arid ecosystems, f e r r o i including deserts, suggested that increases in aridity s F t

t

h may result in a decrease in stability and genetic poten- a 101 w tial of the soil microbiome . Therefore, the question is ↓SOC ↓SOC how resilient the soil microbiome will be to increasing periods of drought and less predictable weather patterns in the future. ↑SOC

S e ↓SOC Increased precipitation and flooding. Some areas are a w a t experiencing increases in soil moisture due to flooding t h e g r u and/or severe and erratic precipitation events. For exam- in o t r ru D ple, increased heavy precipitation events are predicted to si on occur in wet tropical regions102. Climate change is also predicted to shift precipitation in northern regions to H S Inc on 2 reased precipitati more rain at the expense of snow, resulting in reduced CO 2 snow pack and increasing freeze–thaw cycles103. As soil CO2 moisture increases, soil pores become water-filled and anaerobic, thus providing ripe conditions for methano- genesis and denitrification, and the potential for release of CH and N O. Differences in moisture and vegetation Fig. 2 | Soil microbial responses to climate change. A soil microbial community of 4 2 due to changing precipitation patterns can give rise to bacteria, archaea (red and blue) and fungal hyphae (green) in the absence of climate change pressures is depicted in the centre. Examples of climate responses are shown contrasting results with respect to microbial community at the periphery (note that changes in cell colour to orange, green or purple indicate a responses. Therefore, predictive metabolic models are community shift). Increases or decreases in soil organic carbon (SOC) are indicated by needed for more accurate simulations under future 104 up and down white arrows, respectively. Elevated CO2 can result in an increase in carbon climate scenarios . below ground due to increases in plant growth, with corresponding increases in soil Wetlands represent the interface between terrestrial microbial biomass and shifts in community composition; note that, in the long term, and aquatic environments, and are hotspots that are sen- SOC may decompose at a faster rate than it is formed. White lines indicate increased sitive to changes in climate. The major determinants of microbial activity. Increased temperature can result in loss of SOC, shifts in bacterial CH4 concentrations in wetlands include soil temperature, and/or archaeal compositions and decreases in fungal abundance. Permafrost thaw water-table depth and composition of SOC105. Wetland results in a deepening of the seasonally thawed active layer and an increase in microbial areas, such as peatlands, may transition from carbon degradation of SOC. Viruses (not shown) have been detected in thawed permafrost and have been implicated in carbon cycling. Depending on landscape hydrology , sinks to carbon sources in the future, thus aggravating current warming trends. However, when peatlands are methanogens in wetter and more anaerobic areas can generate CH4. Drought can flooded, oxidative decomposition of SOC in peat may be result in less decomposition of SOC, lower microbial biomass and less CO2 production. Surviving bacteria may produce molecules to retain cell turgor (osmolytes) and/or enter inhibited, resulting in a net uptake of carbon. a dormant physiological state (as represented by a change in shape of the red cell). Under A special case is the increase in saltwater-impacted drought, fungal hyphae can be better suited to bridge disconnected soil pores and serve soils associated with sea-level rise. Many coastal areas are as a fungal highway for other microbial cells. Increased precipitation can increase water experiencing saltwater intrusion because of the increas- saturation and anaerobic soil zones. This panel illustrates a case for wetting dry soil when ing global sea level, at a rate of 3.2 ± 0.4 mm per year106. there is a sudden increase in water and nutrient availability , which may cause some cells The intrusion of saltwater into vulnerable coastal soil to burst (as shown by the burst red cell) and serve as a substrate for other cells to become ecosystems introduces salt and sulfate: the latter acts more active (white lines by blue cells indicate increased activity), respire and produce as a terminal electron acceptor and changes the redox CO2 (Birch effect); there can also be some community shifts. Seawater intrusion can also increase soil saturation and anaerobic zones. Furthermore, saltwater can introduce cycling dynamics of the system with resultant increases 107 alternate electron acceptors (for example, sulfate) that can result in community shifts in soil microbial biomass and increases in minerali- (purple cells represent sulfate-reducing bacteria). Fire results in a turnover of soil carbon zation of SOC, resulting in increased levels of CO2 pro- and nitrogen stocks, reduction in microbial biomass, depletion of fungi and some duction (Fig. 2). Therefore, the implications are a future community shifts. net increase in greenhouse gas production through

naTure RevIeWs | MiCrobioloGy volume 18 | January 2020 | 41 Reviews

Albedo increased CO2 emission as seawater levels rise. However, that the proportion of fungi decreased and Actinobacteria 114 The amount of light or different coastal soils respond differently to increased increased following fire . However, because differ- radiation that is reflected from salinity108. With extended periods of flooding, microbial ent groups of bacteria, archaea and fungi are more or a surface. activity was shown to decrease due to resource depletion; less heat sensitive, it is difficult to make generalizations representing a ‘boom and bust’ situation109. The ultimate across different ecosystems and the long-term functional climate impact of rising seawater levels will therefore consequences of these shifts are unknown. depend on the soil microbial community dynamics and Indirect effects of fire can also impact the soil micro- the availability of SOC and electron acceptors that gov- biome. For example, fire-induced changes in plant cover ern the balance between carbon and nutrient storage and have a large impact on the soil microbiome due to their release of greenhouse gases. association with plant roots. Additional indirect effects include increased solar penetration, changes in Increased fire frequency or intensity. Fires are globally due to soil blackening, chemical changes in the soil and increasing in frequency and/or intensity as a result of deposition of alkaline ash and charcoal112. Fire-derived extended, drier fire seasons combined with unsustaina- charcoal is chemically and biologically stable and in ble land management practices. There are several com- one study was shown to maintain activity of nitrify- pounding disturbances associated with fire impacts on ing bacteria in boreal forests following fire117. Another soil microorganisms110. Fire can directly feedback to compounding effect of fire is an increase in soil pH — a 30 climate through release of large amounts of CO2 to major driver of soil microbial diversity and richness . the atmosphere111. Fire intensity and duration impact the Fungal diversity in boreal forests was found to initially characteristics of SOC, with higher fire intensities and increase after fire, presumably due to the increase in pH, longer residence times resulting in greater heat transfer but then declined over time111. When fire is compounded to the underlying soil. When the organic layer of soil with other climate extremes, such as drought, the soil combusts, it can generate further heat. In soils with microbiome may be negatively impacted and less resil- higher moisture content, soil heating can be delayed ient to future disturbances114. However, sweeping gene­ but may kill more microorganisms through pasteuri­ ralizations of fire impacts on soil microorganisms are zation112. Extremely intense fires can also destroy soil complicated because some ecosystems, such as biolog- aggregate structure and reduce soil aeration. Post-fire ical soil crusts in fire-adapted grasslands, are naturally consequences include land degradation and erosion as resilient to low-intensity fires118. soil washes into waterways, thus compounding effects on soil ecosystems. Fire also results in a substantial Soil microbiome manipulation efforts reduction in soil carbon and nitrogen stocks113,114. As a Our increasing awareness of the impacts of climate result, the microbial biomass has been shown to decrease change on the soil microbiome is resulting in an emerg- following a fire event due to a depletion of resources to ing urgency to harness soil microbial capabilities to support microbial growth112 (Fig. 2). mitigate the negative consequences of change. These Several factors are important in determining the interests vary from direct manipulation of soil microbial impact of fire on SOC reserves and the soil micro­ communities to indirect manipulation of their functions biome, including landscape topography (for example, through changes in land management practices or use of mountains compared with flatter, low-lying areas) and inoculants as environmental probiotics (Fig. 3). plant inputs. In boreal mountain regions, a serious con- cern with fire is the potential depletion of soil carbon Carbon sequestration. Atmospheric carbon stocks may from thawing permafrost following fire events. One be reduced by sequestration into stable, non-gaseous study found that following an intensive fire event in forms through biotic and/or abiotic processes. Carbon

the Nome Creek area of Alaska, the soil carbon was enters the soil through assimilation of atmospheric CO2, largely drained along with soil water from the moun- mainly by plants but also by autotrophic soil micro­ tain slope, and the underlying permafrost soil layer organisms. The fraction of photosynthate released to thawed115. This effect was accompanied by a decrease the rhizosphere by plants, either through root exudation, in abundance of most members of the soil micro­ sloughed root cap cells or mycorrhizal fungi, is substan- biome, although some bacteria were enriched post-fire tial (up to 20%)119. Carbon inputs stimulate symbiotic (for example, members of the AD3 candidate phylum). and free-living organisms, which spread the carbon By contrast, in a study of temperate pine-dominated through the soil matrix. Microbial biochemical trans- forests, fire resulted in new inputs into the system, such formations of carbon, and subsequent exchange among as charcoal and plant litter112. communities, cause bioavailable forms of carbon to cycle Direct effects of fire on soil microorganisms include and persist in non-bioavailable forms. The capacity of cell death due to protein denaturation and cell lysis and/or soils to sequester carbon is greater in soils with higher combustion, resulting in a reduction in microbial biodiversity45. This sequestration includes the concerted biomass112. However, a short-term increase in available activities of soil bacteria and fungi to produce carbon nutrients immediately after fire can lead to a short-term polymers that facilitate the formation of soil aggregates increase in microbial activity. One study found increased and to occlude soil carbon in the process. microbial biomass and activity immediately following Avenues being explored for carbon storage include fire treatment, with an increase in relative abundances of mining the untapped biochemical capacity of the soil archaea, and hypothesized that this was because archaeal microbiome for novel reactions that increase the depo­ cell walls are more heat resistant116. Another study found sition of carbon into soil. Single microorganisms or

42 | January 2020 | volume 18 www.nature.com/nrmicro Reviews

photosynthate to optimize the plant–microorganism– soil system for both optimal plant yield and soil carbon deposition41. To avoid field application of genetically modified plants, this will require collaboration between plant breeders and soil microbiologists to design the best pairing of specific beneficial microorganisms with specific plant genotypes125.

Microorganisms as beneficial plant inoculants. Beneficial plant growth-promoting (PGP) bacteria and fungi that inhabit the rhizosphere may help counteract the nega- BacteriaNEPS ecromass Nodule tive consequences of drought by optimizing growth of plants in increasingly stressful conditions126 (Fig. 3). PGP microorganisms can be applied as seed coatings, or as N 2 liquid or granular supplements to plants growing in the field. The classical example of a PGP strain is that of Rhizobium spp. inoculants that are applied for biologi- IAA cal nitrogen fixation in association with legumes (Fig. 3). Currently, there is growing interest in going beyond tra- Water retention Plant Carbon sequestration Plant growth promotion ditional application of inoculants as biofertilizers and Mycorrhizal fungi biopesticides, to also harness other beneficial proper- ties of PGP microorganisms to mitigate the deleterious Fig. 3 | Manipulating the soil microbiome to mitigate the negative consequences of 127 climate change. Examples of methods to harness the soil microbiome to mitigate the consequences of climate change . negative consequences of climate change are shown. Microorganisms can improve water Several avenues are being explored to use PGP 125 retention in soil as a drought mitigation strategy ; for example, through production of microorganisms to alleviate drought stress in plants . extracellular polymeric substances (EPS) to plug soil pores as a strategy for retaining soil For example, some soil bacteria produce extracellular water under dry conditions. Soil microorganisms can serve as a plant carbon sink through polymeric substances, resulting in hydrophobic biofilms microbial uptake of carbon exported from plant roots that is subsequently stored as that can protect plants from desiccation126. Beneficial cellular biomass or transformed to stable metabolites. Carbon can be sequestered in soil soil microorganisms could also be exploited to increase as dead microbial biomass (necromass). Plant growth-promoting microorganisms can be tolerance of crops to drought stress through their pro- used to enhance plant production in soils that are negatively impacted by climate duction of phytohormones that stimulate plant growth, change. Examples include increasing the provision of nutrients such as N through 2 accumulation of osmolytes or other protective com- symbiotic or associative nitrogen-fixing bacteria, enhancing nutrient uptake through 125,128 mycorrhizal fungi and the production of microbial plant growth-promoting hormones, pounds, or detoxification of reactive oxygen species such as indole-3-acetic acid (IAA). (Fig. 3). For example, some bacteria synthesize indole- 3-acetic acid in the rhizosphere, resulting in increased root production129 that can help alleviate water stress125. interacting members of consortia that catalyse these Rhizosphere microorganisms have also been shown to reactions can drive the carbon decomposition pathways secrete metabolites that can accumulate in plant cells towards more recalcitrant and stable end products120. to alleviate osmotic stress130,131. Resistance to drought Alternatively, soil microbiomes can be manipulated stress and nutrient acquisition can also be enhanced in situ, through addition of amendments that enhance by associations with beneficial arbuscular mycorrhizal their activity to take up and store carbon in soil. For fungi, for example, by regulating the plant production example, persistent carbon produced from microbial res- of specific molecules known as aquaporins that reduce idues could be stored in deeper soil layers121. Microbial water stress122,132. Arbuscular mycorrhizal fungi can also residues can include macromolecules from extracellular directly access water by extension of their mycelia into polymeric substances or dead biomass (necromass) (Fig. 3) water-filled soil pores that are not otherwise accessible that have been shown to persist in soil122. Another avenue to plant roots. being considered is to use pyrolysed carbon (biochar) as Synchronizing plant demand with microbial nitro- an amendment to sequester soil carbon in a relatively gen supply also holds great promise for mitigating 123 stable state ; stability depends on whether components microbial N2O production. For example, arbuscular in the biochar are respired by soil microorganisms. mycorrhizal fungi can be used to acquire ammonium

Interactions between plants and soil rhizosphere and mitigate N2O production. Other biological strate-

Necromass microbiomes can also be manipulated to facilitate soil gies to mitigate N2O emissions include inoculation with Residue mixtures of molecules 41,124 133 carbon storage . For example, root exudate depo- N2O-consuming communities or blocking nitrifica- derived from microorganisms, including biomass, intracellular sition can be enhanced by increasing the plant sink tion using biological inhibitors of the ammonia oxida- 134 and extracellular biomolecules/ of carbon to the rhizosphere where it is transformed tion pathway . Together, these examples illustrate how aggregations. into stable metabolites and/or stored in microbial bio- beneficial properties carried out by soil microorganisms mass41. In this scenario, the plant can also be geneti- can be leveraged to help maintain ecosystem services in Biochar cally modified to select for beneficial root-colonizing a changing climate. Fire-derived (pyrolysed) carbon (also known as black carbon) microorganisms that trap specific carbon exudates Agriculture results in additional compounding that has been proposed as a soil produced by the plant (Fig. 3). Future strategies could factors that are not discussed in this Review. It is carbon-storage amendment. lead to an ability to genetically control the allocation of noteworthy that the US National Academy of Science

naTure RevIeWs | MiCrobioloGy volume 18 | January 2020 | 43 Reviews

recently published the scientific breakthroughs needed by soil microbiomes are vital for retaining soil carbon for food and agriculture by 2030 (ref.135), which include and for the provision of nutrients to plants, and the a strategy for manipulating the soil microbiome to importance of soil microbiomes in preserving a healthy increase crop productivity in the face of climate change. soil for future generations cannot be overstated. There Additional research priorities include a need to under- is an urgent need to gain a better understanding of the stand the biogeochemical pathways underpinning SOC repercussions of climate change on key biogeochemi- decomposition and production of greenhouse gases in cal processes carried out by soil microorganisms, to use order to derive better practices to prevent carbon loss this information to make better predictions of climate from soil. For example, although the redox chemistry of impacts and to, ultimately, design microbial strategies electron donor molecules is understood, the energetics to combat further climate warming and soil degrada- and thermodynamics of organic carbon electron accep- tion. Although we review several methods to harness tors that drive microbial metabolism are not well char- soil microorganisms to help mitigate climate change, acterized in the context of the soil habitat. Our present we do not by any means propose that this will be suf- frontier is to describe the physiological response, or meta­ ficient to counterbalance the loss of soil and the gen- phenome, of the soil microbiome. This knowledge will eration of greenhouse gases that is already occurring. facilitate predictions of the impacts of climate change Instead, an integrated approach is urgently needed that on soil functions and better enable harnessing bene­ would employ best practices for sustainable soil man- ficial properties of the soil microbiome to help mitigate agement to support plant production, store and supply negative consequences of climate change. clean water, maintain biodiversity, sequester carbon and increase resilience in a changing climate. To achieve this Conclusions goal, we must connect the fine-scale detail arising from While writing this Review, we came to have an even microbiome studies to the landscape-scale resolution greater appreciation of the value of our living soil and of ecosystem services and many Earth system climate a heightened concern about the future as these fragile models. Most importantly, we need the political will and resources are being imperilled by the negative conse- a global dedicated effort to curb the emissions of green- quences of climate change. Soil, which many ‘take for house gases that are the root of climate change. We note granted’ and often overlook, is a non-renewable resource that these concerns are in line with the recent ‘warning and is currently being depleted at a faster rate than it to humanity’ that serves as a wake-up call to the impor- is being formed136. Although not within the scope of tance of microorganisms in establishing ecosystem this Review, when combined with unsustainable soil stability in the future as climate changes137. management practices, we are on course to lose much of our fertile soils. The ecosystem services carried out Published online 4 October 2019

1. U.S. Global Change Research Program. Impacts, risks, 12. Rustad, L. et al. A meta-analysis of the response of soil 24. Evans, S. E. & Wallenstein, M. D. Climate change and adaptation in the United States: Fourth National respiration, net nitrogen mineralization, and alters ecological strategies of soil bacteria. Ecol. Lett. Climate Assessment Vol. II (USGCRP, 2018). aboveground plant growth to experimental ecosystem 17, 155–164 (2014). 2. Friedlingstein, P. et al. Climate–carbon cycle warming. Oecologia 126, 543–562 (2001). This study assesses the phylogenetic conservation feedback analysis: results from the C4MIP model 13. Nottingham, A. T. et al. Climate warming and soil of ecological strategies in response to drying– intercomparison. J. Clim. 19, 3337–3353 (2006). carbon in tropical forests: insights from an elevation re-wetting in incubation studies on soils from the 3. Wang, K. et al. Modeling global soil carbon and soil gradient in the Peruvian Andes. Bioscience 65, Rainfall Manipulation Plot Study (RaMPS) in microbial carbon by integrating microbial processes 906–921 (2015). the US tallgrass prairie. into the ecosystem process model TRIPLEX-GHG. 14. Vitousek, P. M. et al. Human alteration of the global 25. Berg, M. P. et al. Adapt or disperse: understanding J. Adv. Modeling Earth Syst. 9, 2368–2384 nitrogen cycle: sources and consequences. Ecol. Appl. species persistence in a changing world. Glob. Chang. (2017). 7, 737–750 (1997). Biol. 16, 587–598 (2010). 4. Wieder, W. R., Bonan, G. B. & Allison, S. D. Global soil 15. Syakila, A. & Kroeze, C. The global nitrous oxide 26. de Vries, F. T. & Shade, A. Controls on soil microbial carbon projections are improved by modelling budget revisited. Greenh. Gas Meas. Manag. 1, community stability under climate change. Front. microbial processes. Nat. Clim. Change 3, 909–912 17–26 (2011). Microbiol. 4, 265 (2013). (2013). 16. Regan, K. et al. Spatial and temporal dynamics of 27. Thompson, L. R. et al. A communal catalogue reveals 5. Wahl T. et al. When environmental forces collide. nitrogen fixing, nitrifying and denitrifying microbes in Earth’s multiscale microbial diversity. Nature 551, Eos.org https://eos.org/features/when-environmental- an unfertilized grassland soil. Soil Biol. Biochem. 109, 457–463 (2017). forces-collide (2018). 214–226 (2017). 28. Jansson, J. K. & Hofmockel, K. S. The soil 6. Berard, A., Pierre, R. & Kaisermann, A. Soil microbial 17. Groffman, P. M. et al. Methods for measuring microbiome—from metagenomics to metaphenomics. community responses to heat wave components: denitrification: diverse approaches to a difficult Curr. Opin. Microbiol. 43, 162–168 (2018). drought and high temperature. Clim. Res. 3, problem. Ecol. Appl. 16, 2091–2122 (2006). 29. Waldrop, M. P. & Firestone, M. K. Response of 243–264 (2014). 18. Butterbach-Bahl, K., Baggs Elizabeth, M., microbial community composition and function 7. Sheik, C. et al. Effect of warming and drought on Dannenmann, M., Kiese, R. & Zechmeister- to soil climate change. Microb. Ecol. 52, 716–724 grassland microbial communities. ISME J. 5, Boltenstern, S. Nitrous oxide emissions from soils: (2006). 1692–1700 (2011). how well do we understand the processes and their 30. Fierer, N. & Jackson, R. B. The diversity and 8. Tarnocai, C. et al. Soil organic carbon pools in the controls? Phil. Trans. R. Soc. B 368, 20130122 biogeography of soil bacterial communities. northern circumpolar permafrost region. (2013). Proc. Natl Acad. Sci. USA 103, 626–631 (2006). Glob. Biogeochem. Cycles 23, GB2023 (2009). 19. van Kessel, M. A. H. J. et al. Complete nitrification by 31. Schimel, J. & Schaeffer, S. Microbial control over 9. Bruhwiler, L. et al. in Second State of the Carbon a single microorganism. Nature 528, 555–559 carbon cycling in soil. Front. Microbiol. 3, 348 Cycle Report (SOCCR2): a sustained assessment (2015). (2012). report (eds Cavallaro, N. et al.) 42–70 (U.S. Global 20. Xia, F. et al. Ubiquity and diversity of complete 32. Cordero, O. X. & Datta, M. S. Microbial interactions Change Research Program, 2018). ammonia oxidizers (comammox). Appl. Environ. and community assembly at microscales. Curr. Opin. 10. Bradford, M. A. et al. Thermal adaptation of soil Microb. 84, e01390-18 (2018). Microbiol. 31, 227–234 (2016). microbial respiration to elevated temperature. 21. Kuzyakov, Y. & Blagodatskaya, E. Microbial 33. Dunbar, J. et al. Common bacterial responses in six Ecol. Lett. 11, 1316–1327 (2008). hotspots and hot moments in soil: concept & review. ecosystems exposed to 10 years of elevated 11. Bond-Lamberty, B., Bailey, V., Chen, M., Gough, C. & Soil Biol. Biochem. 83, 184–199 (2015). atmospheric carbon dioxide. Environ. Microbiol. 14, Vargas, R. Globally rising soil heterotrophic respiration 22. Norby, R. J. et al. Model–data synthesis for the next 1145–1158 (2012).

over recent decades. Nature 560, 80–83 (2018). generation of forest free-air CO2 enrichment (FACE) 34. Hayden, H. L. et al. Changes in the microbial Based on existing data, this study determines experiments. New Phytol. 209, 17–28 (2016). community structure of bacteria, archaea and fungi in

trends towards increased soil microbial 23. Schimel, J., Balser, T. C. & Wallenstein, M. Microbial response to elevated CO2 and warming in an

mineralization of SOC, leading to increased CO2 stress-response physiology and its implications for Australian native grassland soil. Environ. Microbiol. emissions as a result of climate change. ecosystem function. Ecology 88, 1386–1394 (2007). 14, 3081–3096 (2012).

44 | January 2020 | volume 18 www.nature.com/nrmicro Reviews

35. Allison, S. D. A trait-based approach for modelling 56. Deslippe, J. R., Hartmann, M., Simard, S. W. & This article reviews current knowledge of microbial microbial litter decomposition. Ecol. Lett. 15, Mohn, W. W. Long-term warming alters the community dynamics and physiological responses 1058–1070 (2012). composition of Arctic soil microbial communities. to drought.

This article describes a trait-based modelling FEMS Microbiol. Ecol. 82, 303–315 (2012). 79. Hyvonen, R. et al. The likely impact of elevated [CO2], approach that links phylogenetic and functional 57. Guo, X. et al. Climate warming leads to divergent nitrogen deposition, increased temperature and information to predict ecosystem processes carried succession of grassland microbial communities. management on carbon sequestration in temperate out by microbial communities. Nat. Clim. Change 8, 813–818 (2018). and boreal forest ecosystems: a literature review. 36. Tu, Q. et al. Metagenomic reconstruction of nitrogen 58. Frey, S. D., Drijber, R., Smith, H. & Melillo, J. New Phytol. 173, 463–480 (2007).

cycling pathways in a CO2-enriched grassland Microbial biomass, functional capacity, and 80. Pointing, S. B. & Belnap, J. Microbial colonization and ecosystem. Soil Biol. Biochem. 106, 99–108 (2017). community structure after 12 years of soil warming. controls in dryland systems. Nat. Rev. Microbiol. 10,

37. Yu, H. et al. Elevated CO2 and warming altered Soil Biol. Biochem. 40, 2904–2907 (2008). 551–562 (2012). grassland microbial communities in soil top-layers. 59. Zhou, J. et al. Microbial mediation of carbon-cycle 81. Garcia-Pichel, F., Johnson, S. L., Youngkin, D. & Front. Microbiol. 9, 1790 (2018). feedbacks to climate warming. Nat. Clim. Change 2, Belnap, J. Small-scale vertical distribution of bacterial This study uses a gene array to measure the 106–110 (2011). biomass and diversity in biological soil crusts from functional gene composition, structure and 60. Heimann, M. & Reichstein, M. Terrestrial ecosystem arid lands in the Colorado Plateau. Microb. Ecol. 46, metabolic potential of soil microbial communities carbon dynamics and climate feedbacks. Nature 451, 312–321 (2003).

under warming, eCO2 and eCO2 with warming 289–292 (2008). 82. Steven, B., Kuske, C. R., Gallegos-Graves, L. V., conditions in a semi-arid grassland. This article provides evidence and discusses Reed, S. C. & Belnap, J. Climate change and physical 38. Adair, C. E., Reich, P. B., Trost, J. J. & Hobbie, S. E. uncertainties about terrestrial ecosystem feedback disturbance manipulations result in distinct biological

Elevated CO2 stimulates grassland soil respiration by with climate change. soil crust communities. Appl. Environ. Microb. 81, increasing carbon inputs rather than by enhancing soil 61. Turetsky, M. R. et al. Permafrost collapse is 7448–7459 (2015). moisture. Glob. Chang. Biol. 17, 3546–3563 (2011). accelerating carbon release. Nature 569, 32–34 83. Reed, S. C. et al. in Biological Soil Crusts: 39. Bréchet, L. M. et al. Distinct responses of soil (2019). An Organizing Princincipal in Drylands (eds Weber, B., respiration to experimental litter manipulation in 62. Mackelprang, R., Saleska, S. R., Jacobsen, C. S., Büdel, B. & Belnap, J.) 451–476 (Springer, 2016). temperate woodland and tropical forest. Ecol. Evol. 8, Jansson, J. K. & Taş, N. Permafrost meta-omics and 84. Upton, R. N., Bach, E. M. & Hofmockel, K. S. 3787–3796 (2018). climate change. Annu. Rev. Earth Planet. Sci. 44, Belowground response of prairie restoration and 40. Bengtson, P., Barker, J. & Grayston, S. J. Evidence of a 439–462 (2016). resiliency to drought. Agric. Ecosyst. Environ. 266, strong coupling between root exudation, C and N 63. Hultman, J. et al. Multi-omics of permafrost, active 122–132 (2018). availability, and stimulated SOM decomposition layer and thermokarst bog soil microbiomes. Nature 85. de Vries, F. T. et al. Soil bacterial networks are less caused by rhizosphere priming effects. Ecol. Evol. 2, 521, 208–212 (2015). stable under drought than fungal networks. 1843–1852 (2012). 64. Tas, N. et al. Landscape topography structures the soil Nat. Commun. 9, 3033 (2018). 41. Jansson, C., Vogel, J., Hazen, S., Brutnell, T. & microbiome in Arctic polygonal tundra. Nat. Commun. This study uses network analyses to show that Mockler, T. Climate-smart crops with enhanced 9, 777 (2018). bacterial networks are less resilient to drought photosynthesis. J. Exp. Bot. 69, 3801–3809 (2018). 65. Woodcroft, B. J. et al. Genome-centric view of carbon than fungal networks, suggesting that bacteria are 42. Qiao, N. et al. Labile carbon retention compensates processing in thawing permafrost. Nature 560, less stable to environmental change than fungi.

for CO2 released by priming in forest soils. Glob. 49–54 (2018). 86. Treseder, K. K., Berlemont, R., Allison, S. D. & Chang. Biol. 20, 1943–1954 (2014). This study assembles thousands of genomes from Martiny, A. C. Drought increases the frequencies 43. van Groenigen, K. J., Qi, X., Osenberg, C. W., Luo, Y. Q. soil metagenomes, corresponding to primarily of fungal functional genes related to carbon and & Hungate, B. A. Faster decomposition under uncultivated and uncharacterized microorganisms, nitrogen acquisition. PLOS ONE 13, e0206441

increased atmospheric CO2 limits soil carbon storage. along a permafrost thaw gradient in Sweden. (2018). Science 344, 508–509 (2014). 66. Johnston, E. R. et al. Responses of tundra soil 87. Carson, J. K. et al. Low pore connectivity increases 44. Drake, J. E. et al. Increases in the flux of carbon microbial communities to half a decade of bacterial diversity in soil. Appl. Environ. Microb. 76, belowground stimulate nitrogen uptake and sustain experimental warming at two critical depths. 3936–3942 (2010). the long-term enhancement of forest productivity Proc. Natl Acad. Sci. USA 116, 15096–15105 88. Dechesne, A., Wang, G., Gülez, G., Or, D. & Smets, B. F.

under elevated CO2. Ecol. Lett. 14, 349–357 (2019). Hydration-controlled bacterial motility and dispersal (2011). 67. Bottos, E. M. et al. Dispersal limitation and on surfaces. Proc. Natl Acad. Sci. USA 107, 45. Lal, R. Soil carbon sequestration impacts on global thermodynamic constraints govern spatial structure of 14369–14372 (2010). climate change and food security. Science 304, permafrost microbial communities. FEMS Microbiol. 89. Guhr, A., Borken, W., Spohn, M. & Matzner, E. 1623–1627 (2004). Ecol. 94, 1–14 (2018). Redistribution of soil water by a saprotrophic fungus This paper discusses different land management 68. Mackelprang, R. et al. Metagenomic analysis of a enhances carbon mineralization. Proc. Natl Acad. strategies to optimize soil carbon sequestration permafrost microbial community reveals a rapid Sci. USA 112, 14647–14651 (2015). through soil microbial activities. response to thaw. Nature 480, 368–371 (2011). 90. Barnard, R. L., Osborne, C. A. & Firestone, M. K. 46. Kuzyakov, Y. Priming effects: interactions between 69. Müller, O. et al. Disentangling the complexity of Responses of soil bacterial and fungal communities to living and dead organic matter. Soil Biol. Biochem. 42, permafrost soil by using high resolution profiling of extreme desiccation and rewetting. ISME J. 7, 1363–1371 (2010). microbial community composition, key functions and 2229–2241 (2013). 47. Scharlemann, J. P. W., Tanner, E. V. J., Hiederer, R. & respiration rates. Environ. Microbiol. 20, 4328–4342 This study finds different responses of potentially Kapos, V. Global soil carbon: understanding and (2018). active bacterial and fungal communities to managing the largest terrestrial carbon pool. 70. Singleton, C. M. et al. Methanotrophy across a natural desiccation and re-wetting across three grassland Carbon Manag. 5, 81–91 (2014). permafrost thaw environment. ISME J. 12, sites in California, United States. 48. Melillo, J. M. et al. Long-term pattern and magnitude 2544–2558 (2018). 91. Boot, C. M., Schaeffer, S. M. & Schimel, J. P. Static of soil carbon feedback to the climate system in a 71. Emerson, J. B. et al. Host-linked soil viral ecology osmolyte concentrations in microbial biomass during warming world. Science 358, 101–104 (2017). along a permafrost thaw gradient. Nat. Microbiol. 3, seasonal drought in a California grassland. Soil Biol. This article describes the results of long-term 870–880 (2018). Biochem. 57, 356–361 (2013). warming field studies at the Harvard Forest 72. Trubl, G. et al. Soil viruses are underexplored players 92. Kakumanu, M. L., Cantrell, C. L. & Williams, M. A. field station on soil carbon flux and how the in ecosystem carbon processing. mSystems 3, Microbial community response to varying magnitudes soil microorganisms acclimate to the warmer soil e00076-18 (2018). of desiccation in soil: a test of the osmolyte conditions. This study characterizes viruses in thawing accumulation hypothesis. Soil Biol. Biochem. 57, 49. Romero-Olivares, A. L., Allison, S. D. & Treseder, K. K. permafrost peatlands and suggests that viruses 644–653 (2013). Soil microbes and their response to experimental can impact the biogeochemistry of their hosts and 93. Meisner, A., Leizeaga, A., Rousk, J. & Bååth, E. Partial warming over time: a meta-analysis of field studies. carbon metabolism. drying accelerates bacterial growth recovery to Soil Biol. Biochem. 107, 32–40 (2017). 73. Penton, C. R. et al. Fungal diversity in permafrost and rewetting. Soil Biol. Biochem. 112, 269–276 (2017). 50. Schindlbacher, A. et al. Experimental warming effects tallgrass prairie soils under experimental warming Based on experiments and other recent results, on the microbial community of a temperate mountain conditions. Appl. Environ. Microb. 79, 7063–7072 this article proposes a framework for microbial forest soil. Soil Biol. Biochem. 43, 1417–1425 (2013). response patterns after drying–re-wetting, in which (2011). 74. Schütte, U. M. E. et al. Effect of permafrost thaw on the harshness of drying determines the response 51. Clemmensen, K. E. et al. Roots and associated fungi plant and soil fungal community in a boreal forest: pattern of bacteria upon re-wetting dried soils. drive long-term carbon sequestration in boreal forest. does fungal community change mediate plant 94. Bouskill, N. J. et al. Belowground response to drought Science 339, 1615–1618 (2013). productivity response? J. Ecol. 107, 1737–1752 in a tropical forest soil. II. Change in microbial function 52. Allison, S. D. & Treseder, K. K. Warming and drying (2019). impacts carbon composition. Front. Microbiol. 7, 323 suppress microbial activity and carbon cycling in 75. Cook, B. I., Ault, T. R. & Smerdon, J. E. Unprecedented (2016). boreal forest soils. Glob. Chang. Biol. 14, 2898–2909 21st century drought risk in the American Southwest 95. Naylor, D., DeGraaf, S., Purdom, E. & Coleman-Derr, D. (2008). and Central Plains. Science Adv. 1, e1400082 (2015). Drought and host selection influence bacterial 53. Morrissey, E. M. et al. Evolutionary history constrains 76. Huang, J. P., Yu, H. P., Guan, X. D., Wang, G. Y. & community dynamics in the grass root microbiome. microbial traits across environmental variation. Guo, R. X. Accelerated dryland expansion under ISME J. 11, 2691–2704 (2017). Nat. Ecol. Evol. 3, 1064–1069 (2019). climate change. Nat. Clim. Change 6, 166–171 (2016). 96. Unger, S., Maguas, C., Pereira, J., David, T. & 54. DeAngelis, K. M. et al. Long-term forest soil warming 77. McHugh, T. A. et al. Climate controls prokaryotic Werner, C. The influence of precipitation pulses on alters microbial communities in temperate forest soils. community composition in desert soils of the soil respiration—assessing the “Birch effect” by stable Front. Microbiol. 6, 104 (2015). southwestern United States. FEMS Microbiol. Ecol. carbon isotopes. Soil Biol. Biochem. 42, 1800–1810 55. Zhang, B. et al. Responses of soil microbial 93, fix116 (2017). (2010). communities to experimental warming in alpine 78. Schimmel, J. P. Life in dry soils: effects of drought on 97. Birch, H. F. The effect of soil drying on humus grasslands on the Qinghai–Tibet Plateau. PLOS ONE soil microbial communities and processes. Ann. Rev. decomposition and nitrogen availability. Plant Soil 10, 8, E103859 (2014). Ecol. Evol. Syst. 49, 409–432 (2018). 9–31 (1958).

naTure RevIeWs | MiCrobioloGy volume 18 | January 2020 | 45 Reviews

98. Blazewicz, S. J., Schwartz, E. & Firestone, M. K. shrublands: immediate changes in soil microbial and Agricultural Research by 2030 (National Growth and death of bacteria and fungi underlie community structure and ecosystem functions. Academies, 2019). rainfall-induced carbon dioxide pulses from seasonally Microb. Ecol. 64, 242–255 (2012). 136. Lal, R. Soil erosion and the global carbon budget. dried soil. Ecology 95, 1162–1172 (2014). 117. DeLuca, T. H. & Sala, A. Frequent fire alters nitrogen Environ. Int. 29, 437–450 (2003). 18 This study uses stable isotope probing with H2 O transformations in ponderosa pine stands of the 137. Cavicchioli, R. et al. Scientists’ warning to humanity: combined with quantitative PCR to determine the inland northwest. Ecology 87, 2511–2522 (2006). microorganisms and climate change. Nat. Rev. Microbiol. dynamics of growing and dying bacterial and fungal 118. Bowker, M. A., Belnap, J., Rosentreter, R. & Graham, B. 17, 569–586 (2019). populations following soil desiccation and re-wetting. Wildfire-resistant biological soil crusts and fire- This work presents a petition for a call to action to 99. Stovicek, A., Kim, M., Or, D. & Gillor, O. Microbial induced loss of soil stability in Palouse prairies, USA. better understand and predict how climate change community response to hydration–desiccation cycles Appl. Soil Ecol. 26, 41–52 (2004). will impact crucial processes currently carried out in desert soil. Sci. Rep. 7, 45735 (2017). 119. Averill, C., Turner, B. L. & Finzi, A. C. Mycorrhiza- by Earth’s microorganisms across a range of 100. Schaeffer, S. M., Homyak, P. M., Boot, C. M., mediated competition between plants and ecosystems. Roux-Michollet, D. & Schimel, J. P. Soil carbon and decomposers drives soil carbon storage. 138. Intergovernmental Panel on Climate Change. Climate nitrogen dynamics throughout the summer drought in Nature 505, 543–545 (2014). Change 2007: Synthesis Report. Contribution of a California annual grassland. Soil Biol. Biochem. 115, 120. Hicks, N. et al. Using prokaryotes for carbon capture Working Groups I, II and III to the Fourth Assessment 54–62 (2017). storage. Trends Biotechnol. 35, 22–32 (2017). Report of the Intergovernmental Panel on Climate Using a long-term field experiment in annual dry 121. Schmidt, M. W. I. et al. Persistence of soil organic Change (eds Core Writing Team, Pachauri, R. K. & grasslands, this study demonstrates a positive matter as an ecosystem property. Nature 478, 49–56 Reisinger, A.) (IPCC, 2007). relationship between drought length, microbial (2011). 139. Schuur, E. A. G. et al. Climate change and the biomass carbon and nitrogen, extractable carbon This article describes a paradigm shift in permafrost carbon feedback. Nature 520, 171–179 and nitrogen, and microbial activity upon re-wetting. understanding the composition of persistent SOCs, (2015). 101. Neilson, J. W. et al. Significant impacts of increasing including determination of microbial cell This article discusses evidence for gradual and aridity on the arid soil microbiome. mSystems 2, macromolecules that persist in soil. sustained from e00195-16 (2017). 122. Kapilan, R., Vaziri, M. & Zwiazek, J. J. Regulation of permafrost as the climate warms. The authors 102. Intergovernmental Panel on Climate Change. Climate aquaporins in plants under stress. Biol. Res. 51, 4 also discuss the need for more integration Change 2014: Synthesis Report. Contribution of (2018). of data with models to improve climate change Working Groups I, II and III to the Fifth Assessment 123. Jansson, C., Tuskan, G. A., Wullschleger, S. D. & predictions. Report of the Intergovernmental Panel on Climate Kalluri, U. C. Phytosequestration: carbon 140. Stocker, T. F. et al. Climate Change 2013. The Physical Change (eds Pachauri, R. & Meyer, L.) (IPCC, 2014). biosequestration by plants and the prospects of Science Basis. Working Group I Contribution to the 103. Sorensen, P. O., Templer, P. H. & Finzi, A. C. genetic engineering. Bioscience 60, 685–696 (2010). Fifth Assessment Report of the Intergovernmental Contrasting effects of winter snowpack and soil frost 124. Wallenstein, M. D. Managing and manipulating the Panel on Climate Change. (Cambridge Univ. Press, on growing season microbial biomass and enzyme rhizosphere microbiome for plant health: a systems 2014). activity in two mixed-hardwood forests. approach. Rhizosphere 3, 230–232 (2017). 141. Llado, S., Lopez-Mondejar, R. & Baldrian, P. Forest Biogeochemistry 128, 141–154 (2016). This article discusses several strategies to harness soil bacteria: diversity, involvement in ecosystem 104. Roy Chowdhury, T. et al. Metaphenomic responses of a soil microorganisms in the plant rhizosphere to processes, and response to global change. Microbiol. native prairie soil microbiome to moisture store soil carbon and help mitigate negative Mol. Biol. Rev. 81, e00063 (2017).

perturbations. mSystems 4, e00061-19 (2019). consequences of rising atmospheric CO2 levels. 142. Kirschbaum, M. U. F. Will changes in soil organic 105. Gedney, N., Cox, P. M. & Huntingford, C. Climate 125. Lakshmanan, V., Ray, P. & Craven, K. D. Toward a carbon act as a positive or negative feedback feedback from wetland methane emissions. Geophys. resilient, functional microbiome: drought on global warming? Biogeochemistry 48, 21–51 Res. Lett. 31, L20503 (2004). tolerance-alleviating microbes for sustainable (2000). 106. Moomaw, W. R. et al. Wetlands in a changing climate: agriculture. Methods Mol. Biol. 1631, 69–84 (2017). 143. Ramankutty, N., Evan, A. T., Monfreda, C. & Foley, J. A. science, policy and management. Wetlands 38, 126. Naylor, D. & Coleman-Derr, D. Drought stress and Farming the planet: 1. Geographic distribution of 183–205 (2018). root-associated bacterial communities. Front. Plant global agricultural lands in the year 2000. Glob. 107. Chambers, L. G., Osborne, T. Z. & Reddy, K. R. Effect Sci. 8, 2223 (2018). Biogeochem. Cycles 22, GB1003 (2008). of salinity-altering pulsing events on soil organic 127. Compant, S., van der Heijden, M. G. A. & Sessitsch, A. 144. Jones, D. L., Nguyen, C. & Finlay, R. D. Carbon flow in the carbon loss along an intertidal wetland gradient: Climate change effects on beneficial plant– rhizosphere: carbon trading at the soil–root interface. a laboratory experiment. Biogeochemistry 115, microorganism interactions. FEMS Microbiol. Ecol. Plant Soil 321, 5–33 (2009). 363–383 (2013). 73, 197–214 (2010). 145. Tian, H. et al. The terrestrial biosphere as a net source 108. Steinmuller, H. E. & Chambers, L. G. Can saltwater 128. Vurukonda, S. S. K. P., Vardharajula, S., Shrivastava, M. of greenhouse gases to the atmosphere. Nature 531, intrusion accelerate nutrient export from freshwater & SkZ, A. Enhancement of drought stress tolerance in 225–228 (2016). wetland soils? An experimental approach. Soil Sci. crops by plant growth promoting rhizobacteria. 146. Zhang, Z. et al. Emerging role of wetland methane Soc. Am. J. 82, 283–29 (2018). Microbiol. Res. 184, 13–24 (2016). emissions in driving 21st century climate change. 109. Sjogaard, K. S., Valdemarsen, T. B. & Treusch, A. H. 129. Armada, E., Azcon, R., Lopez-Castillo, O. M., Proc. Natl Acad. Sci. USA 36, 9647–9652 (2017). Responses of an agricultural soil microbiome to Calvo-Polanco, M. & Ruiz-Lozano, J. M. Autochthonous This study determines the potential sensitivities of flooding with seawater after managed coastal arbuscular mycorrhizal fungi and Bacillus thuringiensis terrestrial wetlands to rising temperatures, and

realignment. Microorganisms 6, 1–18 (2018). from a degraded Mediterranean area can be used emphasizes the importance of including CH4 110. Knelman, E. J., Schmidt, K. S., Garayburu-Caruso, V., to improve physiological traits and performance of feedback from wetlands in climate models. Kumar, S. & Graham, B. E. Multiple, compounding a plant of agronomic interest under drought 147. Makhalanyane, T. P. et al. Microbial ecology of hot disturbances in a forest ecosystem: fire increases conditions. Plant Physiol. Biochem. 90, 64–74 desert edaphic systems. FEMS Microbiol. Rev. 39, susceptibility of soil edaphic properties, bacterial (2015). 203–221 (2015). community structure, and function to change with 130. Pereyra, M. A., García, P., Colabelli, M. N., Barassi, C. A. This article reviews the microbial communities of extreme precipitation event. Soil Syst. 3, 1–16 & Creus, C. M. A better water status in wheat hot desert terrestrial biotopes, the processes that (2019). seedlings induced by Azospirillum under osmotic govern their assembly, the possible effects of 111. Sun, H. et al. Fungal community shifts in structure and stress is related to morphological changes in xylem global climate change on hot desert microbial function across a boreal forest fire chronosequence. vessels of the coleoptile. Appl. Soil Ecol. 53, 94–97 communities and the resulting feedback and Appl. Environ. Microb. 81, 7869–7880 (2015). (2012). directions for future research. 112. Hart, S. C., DeLuca, T., Newman, G., MacKenzie, D. & 131. Casanovas, E. M., Barassi, C. & Sueldo, R. J. Boyle, S. Post-fire vegetative dynamics as drivers Azospirillum inoculation mitigates water stress Acknowledgements of microbial community structure and function effects in maize seedlings. JSTOR 30, 343–350 This research was supported by the US Department of Energy in forest soils. For. Ecol. Manag. 220, 166–184 (2002). Office of Biological and Environmental Research (BER) and is (2005). 132. Quiroga, G., Erice, G., Aroca, R., Chaumont, F. & a contribution to the Scientific Focus Area ‘Phenotypic 113. Nave, L. E., Vance, E. D., Swanston, C. W. & Curtis, P. S. Ruiz-Lozano, J. M. Enhanced drought stress tolerance response of the soil microbiome to environmental perturba- Fire effects on temperate forest soil C and N storage. by the arbuscular mycorrhizal symbiosis in a tions’. Pacific Northwest National Laboratory is operated for Ecol. Appl. 21, 1189–1201 (2011). drought-sensitive maize cultivar is related to a broader the Department of Energy by Battelle Memorial Institute 114. Hinojosa, M. B., Parra, A., Laudicina, V. A. & and differential regulation of host plant aquaporins under contract DE-AC05-76RLO1830. Moreno, J. M. Post-fire soil functionality and microbial than in a drought-tolerant cultivar. Front. Plant Sci. 8, community structure in a mediterranean shrubland 1056 (2017). Author contributions subjected to experimental drought. Sci. Total Environ. 133. Itakura, M. et al. Mitigation of nitrous oxide emissions The authors contributed equally to all aspects of the article. 573, 1178–1189 (2016). from soils by Bradyrhizobium japonicum inoculation. 115. Tas, N. et al. Impact of fire on active layer and Nat. Clim. Change 3, 208–212 (2012). Competing interests permafrost microbial communities and metagenomes 134. Subbarao, G. V. et al. Evidence for biological The authors declare no competing interests. in an upland Alaskan boreal forest. ISME J. 8, nitrification inhibition in Brachiaria pastures. Proc. 1904–1919 (2014). Natl Acad. Sci. USA 106, 17302–17307 (2009). Publisher’s note 116. Goberna, M., García, C., Insam, H., Hernández, M. T. 135. National Academies of Sciences, Engineering and Springer Nature remains neutral with regard to jurisdictional & Verdú, M. Burning fire-prone mediterranean Medicine. Science Breakthroughs to Advance Food claims in published maps and institutional affiliations.

46 | January 2020 | volume 18 www.nature.com/nrmicro