Climate drivers

Climate change is expected to affect aquatic ecosystems strongly by altering the quantity, characteristics, processing, and retention of inputs of terrestrial detritus. The principle climate drivers affecting detrital dynamics in aquatic ecosystems are increases in atmospheric CO₂ and its effects on terrestrial organic matter, changes in precipitation and associated hydrologic impacts, and increases in temperature (Fig. 1, Table 1).

Elevated atmospheric CO₂ can stimulate terrestrial productivity and increase detrital inputs to aquatic systems when plant growth is not nutrient limited (Norby et al. 2010). However, increased atmospheric CO₂ decreases detrital quality by causing production of plant litter that is high in C∶N∶P and unpalatable secondary compounds (Rier et al. 2002, Tuchman et al. 2002, Adams et al. 2003, 2005). Leaching of water-soluble, elevated-CO₂-induced compounds can reduce negative effects on detrital quality and processing (Rier et al. 2005).

Altered patterns of precipitation will make extreme events (e.g., droughts and floods) part of flow regimes, which directly influence water quantity and indirectly influence availability (Lake 2003) and biological processing (Hutchens and Wallace 2002, Tibbets and Molles 2005, Sabo et al. 2008) of detrital resources. Drought reduces physical breakdown rates and slows biological processing (Sangiorgio et al. 2007). Higher discharge will accelerate transport and physical breakdown of organic matter (Paul et al. 2006). High-flow events can scour detritus from stream reaches. When reservoirs occur in river networks, a proportion of organic matter transported downstream will become buried in reservoirs (St. Louis et al. 2000), slowing biological processing.

Long-term temperature records show many instances of stream warming that are correlated with rising air temperatures and other drivers of global change (Kaushal et al. 2010). Recent warming also has been associated with global declines in terrestrial net primary productivity (Zhao and Running 2010), which can reduce detrital inputs to aquatic systems. Increases in air temperature will affect the quantity and timing of terrestrial detrital inputs relative to in situ aquatic production (sensu Mulholland et al. 2009), and a tendency toward earlier leafout with warming temperatures suggests a decrease in seasonal autotrophy in light-limited systems (e.g., deciduous, forested streams; Schwartz et al. 2006). Predicted delays in litterfall associated with increased temperature (Menzel et al. 2006) will reduce light to primary producers and will shift the timing of allochthonous inputs to detritivores. Increased temperatures may promote microbial activity and nutrient sequestration (increasing detrital nutrient content) associated with organic matter and result in greater microbial than metazoan processing (sensu Ferreira et al. 2010, Boyero et al. 2011, Ferreira and Chauvet 2011).

Landuse drivers

Changes in land use have dramatic effects on detrital dynamics in aquatic ecosystems (Allan 2004). Reductions and shifts in watershed vegetation determine what forms of C serve as organic matter in stream systems. Altered riparian and upland forests and increased impervious surfaces and loss of large woody debris will reduce retention of organic matter (Entrekin et al. 2007, 2009). Land uses, such as urbanization, mining, and agriculture, differentially affect the quantity, characteristics, and processing (including retention) of detritus (Fig. 1, Table 2).

Landuse changes, such as deforestation, agricultural intensification, and urban development, reduce vegetated land cover and alter the quantity, composition, and retention of detrital inputs to streams and rivers. Deforestation has the most immediate effects on detrital dynamics in adjacent river networks, and those effects propagate to consumers and ecosystem functions. Effects of overall reductions in forest cover include lower production of higher trophic levels, less use of detrital than autochthonous resources by consumers, and less export of organic matter to downstream assemblages (Wallace et al. 1997, England and Rosemond 2004, Wipfli and Musslewhite 2004). Changes in quantity and quality of detrital inputs and processing rates will depend on the specific changes in vegetation (Eggert and Wallace 2003, Kominoski et al. 2011). Riparian and watershed deforestation are expected to increase light availability and, thus, dependence on autochthonous resources (England and Rosemond 2004).

Watershed urbanization reduces inputs of organic matter to recipient ecosystems, alters channel geomorphology, and decreases large woody debris. These changes decrease retention of detritus (Paul and Meyer 2001, Walsh et al. 2005). Processing rates of organic matter typically are higher in urban than in forested watersheds because of excess nutrients and stimulated microbial processing (Pascoal et al. 2005, Imberger et al. 2008) but can decline because of decreases in macroinvertebrate consumers (Chadwick et al. 2006). Physical fragmentation associated with higher storm runoff in urban watersheds can increase losses of organic matter to downstream reaches (Paul et al. 2006). Inputs of organic matter in urban systems are heavily supplemented by wastewater treatment plants (WWTPs), which contribute large amounts of sewage-derived fine particulate organic matter (FPOM) that stimulates microbial processes (Griffith et al. 2009). However, such inputs can be of low quality to metazoan consumers if they contain contaminants (Rosi-Marshall 2004).

Agricultural landuse changes are among the most widespread globally. The doubling of world food production from 1960–2000 alone has increased use of fertilizers, water, and land area placed into cultivation (Tilman 1999, MEA 2005). In agricultural landscapes, conversion of land for food production reduces terrestrial inputs of organic matter to stream ecosystems (Delong and Brusven 1994), and the magnitude of this reduction in detritus will increase with human population growth and caloric demand (Imhoff et al. 2004). Crop by-products (Griffiths et al. 2009) and litter from grasses and herbs (Menninger and Palmer 2007) enter small streams in agricultural areas, where their processing rates can be relatively rapid. Native types of organic matter may be processed more rapidly in agricultural watersheds because fertilizer runoff increases microbial activity (Paul et al. 2006), and decreases in large woody debris and streambank channelization reduce retention of organic matter (Rosi-Marshall et al. 2007).

Other effects of land use can reduce processing rates. These factors include negative effects of genetically modified crop by-products on detritivore growth rates in agricultural streams (Rosi-Marshall et al. 2007), inhibition of biological processing by acid and heavy metals in mining regions (Dangles et al. 2004, Duarte et al. 2008, Roussel et al. 2008), and general increases in sediment loading that reduce algal production and invertebrate processing of organic matter (Schofield et al. 2004, Pascoal et al. 2005).

Biodiversity drivers

Global patterns of predicted changes in terrestrial vegetation include: 1) changes in timing of litter input caused by shifts in deciduous vs conifer/evergreen species, 2) changes in plant composition via increases in N-fixing and plantation and crop plant species, and 3) reductions in the genetic diversity of riparian vegetation. Changes in the diversity and composition of terrestrial vegetation can affect recipient aquatic ecosystems through alterations in inputs of organic matter (Fig. 1) and resultant in situ dynamics (Table 3). No general pattern linking biodiversity and ecosystem functioning has been agreed upon (Hooper et al. 2005), but the insurance (sensu Hector et al. 1999) afforded aquatic ecosystems by detrital resources with diverse traits could be lost with large-scale changes in terrestrial vegetation, including the expansion of plantation and crop plant species.

Large-scale losses of conifer species are occurring throughout North America because of pathogens, forest harvesting, and fire suppression (Ellison et al. 2005, van Mantgem et al. 2009). Models predict widespread declines in conifers in boreal regions of the northern hemisphere (Cramer et al. 2001). Loss of riparian conifers increases light availability and temperature variation in forested stream ecosystems (Ellison et al. 2005). These losses probably will lead to greater autochthony and increased breakdown rates if conifers are replaced by deciduous tree species because deciduous litter is of higher quality (Richardson et al. 2004).

Additional shifts in watershed vegetation include widespread displacement of native vegetation by crop and plantation plants and N-fixing tree species. Pinus and Eucalyptus plantations produce low-quality (low-nutrient or high-lignin) litter and have partially replaced higher-quality litter inputs of native deciduous riparian species with subsequent negative effects on stream food webs (Graça et al. 2002, Larrañaga et al. 2009). In contrast, increases in higher-quality resources, such as N-fixing deciduous trees (Elaeagnus spp. and Alnus spp.) and fertilized crops in cultivated landscapes, can increase processing rates of organic matter (Griffiths et al. 2009, Follstad Shah et al. 2010, Kominoski et al. 2011).

Increases in species invasions and declines in native and naturally hybridizing riparian tree species will reduce the genetic diversity of riparian plant species (Whitham et al. 2006). Genetic differences in C sequestration (e.g., root structure and stability; Fischer et al. 2007) and leaf phytochemistry (e.g., secondary compounds, such as tannins) among hybridizing plant species (Bailey et al. 2009) can influence the relative inputs and biological processing of organic matter in adjacent stream ecosystems (LeRoy et al. 2006, 2007, Fischer et al. 2007, Bailey et al. 2009, Axelsson et al. 2010). Genetic variation in riparian tree species (e.g., Populus spp.) influences the quality (C∶N∶P) and quantity (some species allocate more C to roots vs leaves) of allochthonous detritus and subsequently alters rates of riparian and stream ecosystem processes (Whitham et al. 2006, Bailey et al. 2009).

Interactive effects on quantity and phenology of organic matter

We predict overall reductions in inputs of terrestrial organic matter caused by declines in forest cover and productivity, especially in tropical rainforests of the southern hemisphere (Zhao and Running 2010) and other vegetated land areas (MEA 2005). Retention of detritus will decrease in stream reaches with increased runoff and channel scour. These reductions may be offset in some areas by increases in dissolved organic C (DOC) associated with increasing temperatures (Evans et al. 2006) and particulate organic C (POC) associated with WWTPs. Increases in autotrophic production will occur in systems with increased light availability, but decreased in-stream autotrophic production is expected throughout the northern hemisphere in systems where earlier seasonal onset and leafout and delayed leaffall of terrestrial vegetation occur (Menzel et al. 2006, Schwartz et al. 2006; but see Mulholland et al. 2009). Changes in spatiolongitudinal patterns of detrital resources along river networks will occur as increasing concentrations of detritus are stored in an increasing number of reservoirs (Table 4).

Changes in detrital quantity may be more important in areas directly affected by landuse changes (e.g., agricultural and urban watersheds), whereas changes in detrital phenology may be more important in areas that are less directly affected by land use (e.g., forested headwaters). For example, a late-spring freeze in southeastern USA caused defoliation in deciduous forests. The resultant increase in light availability increased stream gross primary production, nutrient uptake, and consumer biomass in a system that normally would have been shaded (Roberts et al. 2007, Mulholland et al. 2009). Detrital consumers dependent on the quantity and phenology of organic matter have life-history traits that reflect the availability and timing of inputs of organic matter (Richardson 1991, Eggert and Wallace 2003). Production of these organisms may be reduced by global environmental changes that affect the quantity and phenology of organic matter.

Interactive effects on the characteristics and quality of organic matter

We predict overall increases in detrital quality (lower C∶N∶P) as nutrients are mobilized worldwide. This prediction is contrary to predictions of higher C∶N∶P caused by increased CO₂ (Woodward et al. 2010). Increased quality may be partially balanced by other trends in high C:nutrient detrital inputs that would derive from increased cover of low-quality plantation species. POC as a by-product of agricultural crops is variable in quality and is affected by pesticides and genetic modifications of plant material that may inhibit growth and production of consumers. Lower terrestrial inputs coupled with changes in autotrophy will shift the proportions of allochthonous inputs and autotrophic production (Table 4). In addition, rapid changes in terrestrial species composition (e.g., deciduous vs coniferous/evergreen trees), particularly in high-latitude and high-elevation regions (Hickling et al. 2006, Parmesan 2006), will result in net reductions or qualitative changes in terrestrial detritus with indirect effects on aquatic food webs.

Land cover and climate changes strongly influence the characteristics and quality of organic matter in river networks. Agricultural intensification and conversion of natural forests to forest plantations (Matson et al. 1997, Vitousek et al. 1997, MEA 2005) are important regional drivers affecting inputs of organic matter. Cultivated land comprises ∼50% of land area in the USA (Lubowski et al. 2006) and 24% globally (MEA 2005). Globally, forest plantations consist of Eucalyptus globulus (10%), Pinus spp. (20%), and other coniferous species (31%) (MEA 2005). Landuse effects on detrital quality probably are acute at regional scales, but larger-scale drivers, such as nutrient mobilization and elevated atmospheric CO₂, influence detrital characteristics regardless of watershed land use.

Interactive effects on processing rates of organic matter

We expect faster detrital processing rates as warming and physical abrasion associated with flashy hydrology increase. Slower breakdown rates will occur with entrainment and burial in reservoirs. These effects will be especially prominent in Asia and South America, where rates of reservoir construction are the highest in the world (MEA 2005). Contributions of consumer feeding and nutrient loading to breakdown rates can differ based on litter C quality (Royer 1999). Systems with increased nutrient loading or autotrophic (and associated phytodetritus) production generally will have faster rates of detrital processing, but regions with current and predicted water scarcity are projected to have slower processing rates (Table 4). High-latitude and high-altitude ecosystems are experiencing faster warming than low-latitude and low-altitude ecosystems. This warming has altered permafrost and snowmelt regimes (Hassan et al. 2005). We predict that regions with increased runoff (high latitudes and high altitudes; Milly et al. 2005) will have faster physical processing of organic matter, whereas systems in arid regions with forecasted increased water scarcity will experience slower processing rates (IPCC 2007).

The extent to which increases in temperature and nutrient mobilization affect detrital processing depends upon detrital resource quality and aquatic foodweb structure. For example, E. globulus litter decomposes more slowly than native plant litter in streams because it contains oils and other constituents that inhibit fungal colonization and subsequent consumption and growth by macroinvertebrate shredders (Graça et al. 2002, Larrañaga et al. 2009). In comparison, litter from heavily fertilized crops has lower C∶N∶P than natural riparian vegetation and is processed more rapidly than litter from unfertilized sources (Griffiths et al. 2009). Some investigators have found efficient transfer of energy from basal resources to consumers in nutrient-enriched systems (Slavik et al. 2004), whereas others have identified trophic dead ends caused by inedible resources (e.g., production of detritivorous prey that constrain gape-limited predators; Davis et al. 2010). Decomposition rates may be elevated in systems with increased temperatures and nutrients (urban, agricultural streams). However, these systems also tend to have lower densities of shredding invertebrates (Hagen et al. 2006) and other abiotic stressors that can reduce organic-matter processing rates (Paul and Meyer 2001, Walsh et al. 2005).

Reach-scale effects of organic matter changes on freshwater consumers

At smaller spatial scales (those that occur from molecular to reach scales), effects of global change on organic matter probably will affect balances of energy flow to different consumers (inset circles, Fig. 2). Elevated nutrients stimulate microbial production associated with detritus (Suberkropp et al. 2010) and increase litter breakdown rates (Pascoal et al. 2005, Ferreira et al. 2006, Greenwood et al. 2007). Increased temperatures will stimulate microbial respiration rates (Bärlocher et al. 2008) but may inhibit metazoan feeding rates (Friberg et al. 2009, Ferreira et al. 2010, Boyero et al. 2011). In addition, increases in temperature will select for smaller body sizes among individuals, populations, and communities (Daufresne et al. 2009). Thus, microorganisms may have a selective advantage over larger organisms in warmer environments. Lower quantity and retention and higher quality of detritus coupled with increased temperatures and extreme hydrologic events that increase detrital exports will decrease organic-matter standing crops that support metazoans (Entrekin et al. 2007, 2009), shifting the relative quantity and rates of energy and nutrient flow from organic matter to microbial rather than metazoan consumers.

Increases in extreme hydrologic events (floods and droughts) and decreases in retentive structures (e.g., large woody debris) from deforestation and land clearing will reduce the amount of terrestrial detritus retained in stream ecosystems and support less metazoan secondary production (Entrekin et al. 2007, 2009). Therefore, we predict that detritus will be rapidly exported downstream during floods or will be retained outside of river margins during droughts. Reductions in quantity and quality of terrestrial detritus (e.g., via forest harvesting and crop and plantation species) will lead to declines in secondary production of metazoans, especially in food webs dominated by detritivores (Wallace et al. 1999, Graça et al. 2002, Larrañaga et al. 2009).

Network-scale effects and global C cycling

Processing and retention of organic matter and our ability to predict scaled-up effects of drivers of global change on both will depend on the type and magnitude of effects among longitudinally connected reaches within networks and on the relative strengths of local vs regional drivers. The quantity and quality of detrital inputs change in a downstream direction in river networks (headwaters to mainstems). In many systems, these changes include a shift from coarse particulate organic matter (CPOM) in headwaters to DOC and FPOM in higher-order rivers (Vannote et al. 1980). Changes in CPOM inputs and associated microbial assemblages (fungi; Findlay et al. 2002) will have stronger effects in small, headwater reaches than in larger rivers, and changes in FPOM and DOC and associated microbial assemblages (bacteria; Findlay et al. 2002) will be more important in larger rivers than in headwaters. The extent to which upland and lowland reaches are affected by global changes will determine the cumulative effects on the larger network. For example, effects of global changes will be more predictable and scalable in river networks in which both local (e.g., land use) and regional (e.g., climate) drivers affect terrestrial inputs along the longitudinal gradient of the river network than in river networks with mostly downstream disturbances (e.g., lowland agricultural or urban land use and reservoirs), which will generate more heterogeneous detritus-processing patterns. The spatial heterogeneity of drivers and their elicited effects along longitudinal flow paths must be understood to forecast and quantify network-scale, cumulative effects of global environmental changes on detrital dynamics.

Environmental changes will reallocate detritus along river networks, thereby altering its fates in terms of evasion to the atmosphere and downstream delivery to oceans (Fig. 2). Altered hydrology (droughts and storms) will alter spiraling lengths of C and determine its rate of movement downstream. At network scales, reduced loading and standing crops of terrestrial organic matter and broad-scale alterations in hydrology will disproportionately increase storage of organic matter in reservoirs. Dams and associated reservoirs act as powerful retention structures in river networks by burying organic matter in sediments (Vörösmarty et al. 2003). Global increases in dam building (St. Louis et al. 2000) will increase the proportion of detritus stored in reservoirs (Vörösmarty et al. 2003) and will stimulate sediment respiration and methanogenesis (Tranvik et al. 2009). The flux of nutrients from rivers to oceans has increased with agricultural and industrial expansion (MEA 2005), but retention of detritus and accumulation of nutrients in reservoirs will decrease organic-matter subsidies from rivers to estuaries and coastal ecosystems. Burial of organic matter in reservoirs will decrease its availability to downstream benthic and pelagic food webs and probably will decrease secondary production of consumers (sensu Power et al. 1996, Wallace et al. 1997), increase microbial productivity, and increase C losses through CO₂ and CH₄ evasion (Fig. 2; Tranvik et al. 2009, Williamson et al. 2009). Decreased retention of detritus in upland reaches and increased losses of organic matter to reservoirs could increase C limitation and nutrient accrual in lower-elevation aquatic ecosystems and coastal estuaries (Taylor and Townsend 2010).

Syntheses of the effects of global change on detrital processing from reach-to-network scale will enable predictions of the effects of global change on C-cycling dynamics in inland aquatic systems across continents. Recent syntheses have improved our understanding of the importance of inland waters to the global C cycle (Battin et al. 2008, Downing et al. 2008, Tranvik et al. 2009, Butman and Raymond 2011). Estimates indicate that <50% of the total C that enters inland aquatic ecosystems is exported to the oceans, ∼10% is stored within systems, and >40% is respired to the atmosphere (Cole et al. 2007). We focused on sources and fates of organic C, but terrestrially derived organic C also is transformed and processed in aquatic systems as dissolved inorganic C (DIC) that was respired in soils. These forms of C can be influenced by global change (Jones et al. 2003, Butman and Raymond 2011). Thus, our view of the fate of all terrestrially derived C ultimately must include DIC-mediated pathways. Scientists are beginning to piece together the role of inland aquatic ecosystems in the global C cycle, but environmental changes will continue to alter terrestrial subsidies of organic matter to aquatic ecosystems and the fate of that C in terms of storage, evasion, and transport. We predict that reduced inputs of terrestrial C, a greater proportion of which is retained in reservoirs and processed by microbes as temperature and nutrients increase, will reduce C export to oceans (Fig. 2).

Implications for management

Environmental policies in the USA (e.g., Endangered Species Act, Clean Water Reclamation Act), European Union (e.g., Water Framework Directive), and Australia (Water Act and Water Amendment Acts) focus on water scarcity, pollutants, and biotic integrity. For the most part, these policies fail to address basal resources in their guidance on conservation and management of aquatic biodiversity and ecosystem services. Protecting aquatic ecosystems requires managing watershed attributes that influence water quality, geomorphology, and hydrology (Harding et al. 1998). Broader protection of aquatic ecosystems (e.g., freshwater ecosystem reserves; Abell et al. 2007) will require attention to managing detritus and primary producers as food and habitat resources. Detrital resources are being altered and depleted from aquatic ecosystems with predicted effects on energy flow to higher trophic levels, and recent analyses illustrate the need to consider basal resources when managing populations and ecosystems. For example, managing the effects of pesticides on salmon populations requires attention to effects of pesticides on basal resources (algal and invertebrate food resources) that ultimately affect salmon (Macneale et al. 2010). Assessment of basal food resources could be incorporated explicitly into management practices. For example, indices of detrital integrity (IDIs) could be developed as analogs to indices of biotic integrity (IBIs) with the goal of supporting and maintaining targeted levels of detritus-based biodiversity and ecosystem functioning in aquatic ecosystems.

An essential first step toward developing conservation frameworks that use a bottom-up approach is to quantify relationships among basal resources and ecological outcomes. Frameworks, such as the Ecological Limits of Hydrologic Alteration (ELOHA; Poff et al. 2010), that guide research and management promote development of quantitative relationships between specific system alterations and ecological responses. ELOHA models relationships between flow alterations and ecological outcomes. Models must be developed that address the relationships between modifications of detritus (quantity, quality, and processing rates) and ecological responses. Very few such models currently exist, but a bottom-up approach to management has been recognized as valuable and is being incorporated into programs like the European Union's RIVFUNCTION ( http://www.ecolab.ups-tlse.fr/rivfunction/), which uses litter breakdown rates as a functional response to assess stream condition. We recommend taking steps to develop quantitative models describing the relationships between altered detrital organic matter (a fundamental basal resource in many aquatic ecosystems) and the ecosystem services it provides.

What we need to know: basic to applied research questions

Our review suggests that large changes in inputs and processing of organic matter will occur in aquatic ecosystems. However, we lack quantitative descriptions of relationships between detritus and ecological outcomes, and we have limited understanding of the primary controls and drivers of detritus-based pathways. We recommend undertaking research focused on: 1) quantifying and scaling effects of global change on detritus and 2) determining effects of global change on detritus-based foodweb pathways. In the 1^st case, research should address effects of global change on detrital dynamics (processing, retention, transport) and scaling these effects to river networks. In the 2^nd case, research should address how the effects of global change on detritus propagate through food webs to affect multiple levels of biological organization.