The biological controls of soil carbon accumulation following wildfire and harvest in boreal forests: A review.

Boreal forests are frequently subjected to disturbances, including wildfire and clear-cutting. While these disturbances can cause soil carbon (C) losses, the long-term accumulation dynamics of soil C stocks during subsequent stand development is controlled by biological processes related to the balance of net primary production (NPP) and outputs via heterotrophic respiration and leaching, many of which remain poorly understood. We review the biological processes suggested to influence soil C accumulation in boreal forests. Our review indicates that median C accumulation rates following wildfire and clear-cutting are similar (0.15 and 0.20 Mg ha-1 year-1, respectively), however, variation between studies is extremely high. Further, while many individual studies show linear increases in soil C stocks through time after disturbance, there are indications that C stock recovery is fastest early to mid-succession (e.g. 15-80 years) and then slows as forests mature (e.g. >100 years). We indicate that the rapid build-up of soil C in younger stands appears not only driven by higher plant production, but also by a high rate of mycorrhizal hyphal production, and mycorrhizal suppression of saprotrophs. As stands mature, the balance between reductions in plant and mycorrhizal production, increasing plant litter recalcitrance, and ectomycorrhizal decomposers and saprotrophs have been highlighted as key controls on soil C accumulation rates. While some of these controls appear well understood (e.g. temporal patterns in NPP, changes in aboveground litter quality), many others remain research frontiers. Notably, very little data exists describing and comparing successional patterns of root production, mycorrhizal functional traits, mycorrhizal-saprotroph interactions, or C outputs via heterotrophic respiration and dissolved organic C following different disturbances. We argue that these less frequently described controls require attention, as they will be key not only for understanding ecosystem C balances, but also for representing these dynamics more accurately in soil organic C and Earth system models.

Mineral Soils Are an Important Intermediate Storage Pool of Black Carbon in Fennoscandian Boreal Forests.

Approximately 40% of earth's carbon (C) stored in land vegetation and soil is within the boreal region. This large C pool is subjected to substantial removals and transformations during periodic wildfire. Fire-altered C, commonly known as pyrogenic carbon (PyC), plays a significant role in forest ecosystem functioning and composes a considerable fraction of C transport to limnic and oceanic sediments. While PyC stores are beginning to be quantified globally, knowledge is lacking regarding the drivers of their production and transport across ecosystems. This study used the chemo-thermal oxidation at 375°C (CTO-375) method to isolate a particularly refractory subset of PyC compounds, here called black carbon (BC), finding an average increase of 11.6 g BC m-2 at 1 year postfire in 50 separate wildfires occurring in Sweden during 2018. These increases could not be linked to proposed drivers, however BC storage in 50 additional nearby unburnt soils related strongly to soil mass while its proportion of the larger C pool related negatively to soil C:N. Fire approximately doubled BC stocks in the mineral layer but had no significant effect on BC in the organic layer where it was likely produced. Suppressed decomposition rates and low heating during fire in mineral subsoil relative to upper layers suggests potential removals of the doubled mineral layer BC are more likely transported out of the soil system than degraded in situ. Therefore, mineral soils are suggested to be an important storage pool for BC that can buffer short-term (production in fire) and long-term (cross-ecosystem transport) BC cycling.

Reindeer control over subarctic treeline alters soil fungal communities with potential consequences for soil carbon storage.

The climate-driven encroachment of shrubs into the Arctic is accompanied by shifts in soil fungal communities that could contribute to a net release of carbon from tundra soils. At the same time, arctic grazers are known to prevent the establishment of deciduous shrubs and, under certain conditions, promote the dominance of evergreen shrubs. As these different vegetation types associate with contrasting fungal communities, the belowground consequences of climate change could vary among grazing regimes. Yet, at present, the impact of grazing on soil fungal communities and their links to soil carbon have remained speculative. Here we tested how soil fungal community composition, diversity and function depend on tree vicinity and long-term reindeer grazing regime and assessed how the fungal communities relate to organic soil carbon stocks in an alpine treeline ecotone in Northern Scandinavia. We determined soil carbon stocks and characterized soil fungal communities directly underneath and >3 m away from mountain birches (Betula pubescens ssp. czerepanovii) in two adjacent 55-year-old grazing regimes with or without summer grazing by reindeer (Rangifer tarandus). We show that the area exposed to year-round grazing dominated by evergreen dwarf shrubs had higher soil C:N ratio, higher fungal abundance and lower fungal diversity compared with the area with only winter grazing and higher abundance of mountain birch. Although soil carbon stocks did not differ between the grazing regimes, stocks were positively associated with root-associated ascomycetes, typical to the year-round grazing regime, and negatively associated with free-living saprotrophs, typical to the winter grazing regime. These findings suggest that when grazers promote dominance of evergreen dwarf shrubs, they induce shifts in soil fungal communities that increase soil carbon sequestration in the long term. Thus, to predict climate-driven changes in soil carbon, grazer-induced shifts in vegetation and soil fungal communities need to be accounted for.

Fine root dynamics across pantropical rainforest ecosystems.

Fine roots constitute a significant component of the net primary productivity (NPP) of forest ecosystems but are much less studied than aboveground NPP. Comparisons across sites and regions are also hampered by inconsistent methodologies, especially in tropical areas. Here, we present a novel dataset of fine root biomass, productivity, residence time, and allocation in tropical old-growth rainforest sites worldwide, measured using consistent methods, and examine how these variables are related to consistently determined soil and climatic characteristics. Our pantropical dataset spans intensive monitoring plots in lowland (wet, semi-deciduous, and deciduous) and montane tropical forests in South America, Africa, and Southeast Asia (n = 47). Large spatial variation in fine root dynamics was observed across montane and lowland forest types. In lowland forests, we found a strong positive linear relationship between fine root productivity and sand content, this relationship was even stronger when we considered the fractional allocation of total NPP to fine roots, demonstrating that understanding allocation adds explanatory power to understanding fine root productivity and total NPP. Fine root residence time was a function of multiple factors: soil sand content, soil pH, and maximum water deficit, with longest residence times in acidic, sandy, and water-stressed soils. In tropical montane forests, on the other hand, a different set of relationships prevailed, highlighting the very different nature of montane and lowland forest biomes. Root productivity was a strong positive linear function of mean annual temperature, root residence time was a strong positive function of soil nitrogen content in montane forests, and lastly decreasing soil P content increased allocation of productivity to fine roots. In contrast to the lowlands, environmental conditions were a better predictor for fine root productivity than for fractional allocation of total NPP to fine roots, suggesting that root productivity is a particularly strong driver of NPP allocation in tropical mountain regions.

Effects of moisture dynamics on bryophyte carbon fluxes in a tropical cloud forest.

Bryophytes play key roles in the ecological function of a number of major world biomes but remain understudied compared with vascular plants. Little is known about bryophyte responses to different aspects of predicted changes in moisture dynamics with climate change. In this study, CO2 fluxes and photosynthetic light responses were measured within bryophyte mesocosms, being subjected to different amounts, frequencies, and types (mist or rainfall) of water addition, both before and after different periods of complete desiccation. Bryophyte carbon fluxes and photosynthetic light response were generally affected by the magnitude and type, but not frequency, of watering events. Desiccation suppressed bryophyte carbon uptake even after rehydration, and the degree of uptake suppression progressively increased with desiccation duration. Estimated ecosystem-level bryophyte respiration and net carbon uptake were c. 58% and c. 3%, respectively, of corresponding fluxes from tree foliage at the site. Our results suggest that a simplified representation of precipitation processes may be sufficient to accurately model bryophyte carbon cycling under future climate scenarios. Further, we find that projected increases in drought could have strong negative impacts on bryophyte and ecosystem carbon storage, with major consequences for a wide range of ecosystem processes.

Impacts of fire on sources of soil CO2 efflux in a dry Amazon rain forest.

Fire at the dry southern margin of the Amazon rainforest could have major consequences for regional soil carbon (C) storage and ecosystem carbon dioxide (CO2 ) emissions, but relatively little information exists about impacts of fire on soil C cycling within this sensitive ecotone. We measured CO2 effluxes from different soil components (ground surface litter, roots, mycorrhizae, soil organic matter) at a large-scale burn experiment designed to simulate a severe but realistic potential future scenario for the region (Fire plot) in Mato Grosso, Brazil, over 1 year, and compared these measurements to replicated data from a nearby, unmodified Control plot. After four burns over 5 years, soil CO2 efflux (Rs ) was ~5.5 t C ha-1  year-1 lower on the Fire plot compared to the Control. Most of the Fire plot Rs reduction was specifically due to lower ground surface litter and root respiration. Mycorrhizal respiration on both plots was around ~20% of Rs . Soil surface temperature appeared to be more important than moisture as a driver of seasonal patterns in Rs at the site. Regular fire events decreased the seasonality of Rs at the study site, due to apparent differences in environmental sensitivities among biotic and abiotic soil components. These findings may contribute toward improved predictions of the amount and temporal pattern of C emissions across the large areas of tropical forest facing increasing fire disturbances associated with climate change and human activities.

The variation of productivity and its allocation along a tropical elevation gradient: a whole carbon budget perspective.

Why do forest productivity and biomass decline with elevation? To address this question, research to date generally has focused on correlative approaches describing changes in woody growth and biomass with elevation. We present a novel, mechanistic approach to this question by quantifying the autotrophic carbon budget in 16 forest plots along a 3300 m elevation transect in Peru. Low growth rates at high elevations appear primarily driven by low gross primary productivity (GPP), with little shift in either carbon use efficiency (CUE) or allocation of net primary productivity (NPP) between wood, fine roots and canopy. The lack of trend in CUE implies that the proportion of photosynthate allocated to autotrophic respiration is not sensitive to temperature. Rather than a gradual linear decline in productivity, there is some limited but nonconclusive evidence of a sharp transition in NPP between submontane and montane forests, which may be caused by cloud immersion effects within the cloud forest zone. Leaf-level photosynthetic parameters do not decline with elevation, implying that nutrient limitation does not restrict photosynthesis at high elevations. Our data demonstrate the potential of whole carbon budget perspectives to provide a deeper understanding of controls on ecosystem functioning and carbon cycling.

Informing climate models with rapid chamber measurements of forest carbon uptake.

Models predicting ecosystem carbon dioxide (CO2 ) exchange under future climate change rely on relatively few real-world tests of their assumptions and outputs. Here, we demonstrate a rapid and cost-effective method to estimate CO2 exchange from intact vegetation patches under varying atmospheric CO2 concentrations. We find that net ecosystem CO2 uptake (NEE) in a boreal forest rose linearly by 4.7 ± 0.2% of the current ambient rate for every 10 ppm CO2 increase, with no detectable influence of foliar biomass, season, or nitrogen (N) fertilization. The lack of any clear short-term NEE response to fertilization in such an N-limited system is inconsistent with the instantaneous downregulation of photosynthesis formalized in many global models. Incorporating an alternative mechanism with considerable empirical support - diversion of excess carbon to storage compounds - into an existing earth system model brings the model output into closer agreement with our field measurements. A global simulation incorporating this modified model reduces a long-standing mismatch between the modeled and observed seasonal amplitude of atmospheric CO2 . Wider application of this chamber approach would provide critical data needed to further improve modeled projections of biosphere-atmosphere CO2 exchange in a changing climate.

Nutrient fluxes from insect herbivory increase during ecosystem retrogression in boreal forest.

Ecological theory, developed largely from ungulates and grassland systems, predicts that herbivory accelerates nutrient cycling more in productive than unproductive systems. This prediction may be important for understanding patterns of ecosystem change over time and space, but its applicability to other ecosystems and types of herbivore remain uncertain. We estimated fluxes of nitrogen (N) and phosphorus (P) from herbivory of a common tree species (Betula pubescens) by a common species of herbivorous insect along a -5000-yr boreal chronosequence. Contrary to established theory, fluxes of N and P via herbivory increased along the chronosequence despite a decline in plant productivity. The herbivore-mediated N and P fluxes to the soil are comparable to the main alternative pathway for these nutrients via tree leaf litterfall. We conclude that insect herbivores can make large contributions to nutrient cycling even in unproductive systems, and influence the rate and pattern of ecosystem development, particularly in systems with low external nutrient inputs.

Greater carbon allocation to mycorrhizal fungi reduces tree nitrogen uptake in a boreal forest.

The central role that ectomycorrhizal (EM) symbioses play in the structure and function of boreal forests pivots around the common assumption that carbon (C) and nitrogen (N) are exchanged at rates favorable for plant growth. However, this may not always be the case. It has been hypothesized that the benefits mycorrhizal fungi convey to their host plants strongly depends upon the availability of C and N, both of which are rapidly changing as a result of intensified human land use and climate change. Using large-scale shading and N addition treatments, we assessed the independent and interactive effects of changes in C and N supply on the transfer of N in intact EM associations with -15 yr. old Scots pine trees. To assess the dynamics of N transfer in EM symbioses, we added trace amounts of highly enriched 5NO3(-) label to the EM-dominated mor-layer and followed the fate of the 15N label in tree foliage, fungal chitin on EM root tips, and EM sporocarps. Despite no change in leaf biomass, shading resulted in reduced tree C uptake, ca. 40% lower fungal biomass on EM root tips, and greater 15N label in tree foliage compared to unshaded control plots, where more 15N label was found in fungal biomass on EM colonized root tips. Short-term addition of N shifted the incorporation of 15N label from EM fungi to tree foliage, despite no significant changes in below-ground tree C allocation to EM fungi. Contrary to the common assumption that C and N are exchanged at rates favorable for plant growth, our results show for the first time that under N-limited conditions greater C allocation to EM fungi in the field results in reduced, not increased, N transfer to host trees. Moreover, given the ubiquitous nature of mycorrhizal symbioses, our results stress the need to incorporate mycorrhizal dynamics into process-based ecosystem models to better predict forest C and N cycles in light of global climate change.

The linkages between photosynthesis, productivity, growth and biomass in lowland Amazonian forests.

Understanding the relationship between photosynthesis, net primary productivity and growth in forest ecosystems is key to understanding how these ecosystems will respond to global anthropogenic change, yet the linkages among these components are rarely explored in detail. We provide the first comprehensive description of the productivity, respiration and carbon allocation of contrasting lowland Amazonian forests spanning gradients in seasonal water deficit and soil fertility. Using the largest data set assembled to date, ten sites in three countries all studied with a standardized methodology, we find that (i) gross primary productivity (GPP) has a simple relationship with seasonal water deficit, but that (ii) site-to-site variations in GPP have little power in explaining site-to-site spatial variations in net primary productivity (NPP) or growth because of concomitant changes in carbon use efficiency (CUE), and conversely, the woody growth rate of a tropical forest is a very poor proxy for its productivity. Moreover, (iii) spatial patterns of biomass are much more driven by patterns of residence times (i.e. tree mortality rates) than by spatial variation in productivity or tree growth. Current theory and models of tropical forest carbon cycling under projected scenarios of global atmospheric change can benefit from advancing beyond a focus on GPP. By improving our understanding of poorly understood processes such as CUE, NPP allocation and biomass turnover times, we can provide more complete and mechanistic approaches to linking climate and tropical forest carbon cycling.

Herbivory makes major contributions to ecosystem carbon and nutrient cycling in tropical forests.

The functional role of herbivores in tropical rainforests remains poorly understood. We quantified the magnitude of, and underlying controls on, carbon, nitrogen and phosphorus cycled by invertebrate herbivory along a 2800 m elevational gradient in the tropical Andes spanning 12°C mean annual temperature. We find, firstly, that leaf area loss is greater at warmer sites with lower foliar phosphorus, and secondly, that the estimated herbivore-mediated flux of foliar nitrogen and phosphorus from plants to soil via leaf area loss is similar to, or greater than, other major sources of these nutrients in tropical forests. Finally, we estimate that herbivores consume a significant portion of plant carbon, potentially causing major shifts in the pattern of plant and soil carbon cycling. We conclude that future shifts in herbivore abundance and activity as a result of environmental change could have major impacts on soil fertility and ecosystem carbon sequestration in tropical forests.

Allocation trade-offs dominate the response of tropical forest growth to seasonal and interannual drought.

What determines the seasonal and interannual variation of growth rates in trees in a tropical forest? We explore this question with a novel four-year high-temporal-resolution data set of carbon allocation from two forest plots in the Bolivian Amazon. The forests show strong seasonal variation in tree wood growth rates, which are largely explained by shifts in carbon allocation, and not by shifts in total productivity. At the deeper soil plot, there was a clear seasonal trade-off between wood and canopy NPP, while the shallower soils plot showed a contrasting seasonal trade-off between wood and fine roots. Although a strong 2010 drought reduced photosynthesis, NPP remained constant and increased in the six-month period following the drought, which indicates usage of significant nonstructural carbohydrate stores. Following the drought, carbon allocation increased initially towards the canopy, and then in the following year, allocation increased towards fine-root production. Had we only measured woody growth at these sites and inferred total NPP, we would have misinterpreted both the seasonal and interannual responses. In many tropical forest ecosystems, we propose that changing tree growth rates are more likely to reflect shifts in allocation rather than changes in overall productivity. Only a whole NPP allocation perspective can correctly interpret the relationship between changes in growth and changes in productivity.

Nutrient limitation in rainforests and cloud forests along a 3,000-m elevation gradient in the Peruvian Andes.

We report results from a large-scale nutrient fertilization experiment along a "megadiverse" (154 unique species were included in the study) 3,000-m elevation transect in the Peruvian Andes and adjacent lowland Amazonia. Our objectives were to test if nitrogen (N) and phosphorus (P) limitation shift along this elevation gradient, and to determine how an alleviation of nutrient limitation would manifest in ecosystem changes. Tree height decreased with increasing elevation, but leaf area index (LAI) and diameter at breast height (DBH) did not vary with elevation. Leaf N:P decreased with increasing elevation (from 24 at 200 m to 11 at 3,000 m), suggesting increased N limitation and decreased P limitation with increasing elevation. After 4 years of fertilization (N, P, N + P), plots at the lowland site (200 m) fertilized with N + P showed greater relative growth rates in DBH than did the control plots; no significant differences were evident at the 1,000 m site, and plots fertilized with N at the highest elevation sites (1,500, 3,000 m) showed greater relative growth rates in DBH than did the control plots, again suggesting increased N constraint with elevation. Across elevations in general N fertilization led to an increase in microbial respiration, while P and N + P addition led to an increase in root respiration and corresponding decrease in hyphal respiration. There was no significant canopy response (LAI, leaf nutrients) to fertilization, suggesting that photosynthetic capacity was not N or P limited in these ecosystems. In sum, our study significantly advances ecological understanding of nutrient cycling and ecosystem response in a region where our collective knowledge and data are sparse: we demonstrate N limitation in high elevation tropical montane forests, N and P co-limitation in lowland Amazonia, and a nutrient limitation response manifested not in canopy changes, but rather in stem and belowground changes.

Do Nitrogen and Phosphorus Additions Affect Nitrogen Fixation Associated with Tropical Mosses?

Tropical cloud forests are characterized by abundant and biodiverse mosses which grow epiphytically as well as on the ground. Nitrogen (N)-fixing cyanobacteria live in association with most mosses, and contribute greatly to the N pool via biological nitrogen fixation (BNF). However, the availability of nutrients, especially N and phosphorus (P), can influence BNF rates drastically. To evaluate the effects of increased N and P availability on BNF in mosses, we conducted a laboratory experiment where we added N and P, in isolation and combined, to three mosses (Campylopus sp., Dicranum sp. and Thuidium peruvianum) collected from a cloud forest in Peru. Our results show that N addition almost completely inhibited BNF within a day, whereas P addition caused variable results across moss species. Low N2 fixation rates were observed in Campylopus sp. across the experiment. BNF in Dicranum sp. was decreased by all nutrients, while P additions seemed to promote BNF in T. peruvianum. Hence, each of the three mosses contributes distinctively to the ecosystem N pool depending on nutrient availability. Moreover, increased N input will likely significantly decrease BNF associated with mosses also in tropical cloud forests, thereby limiting N input to these ecosystems via the moss-cyanobacteria pathway.

Ecophysiological controls on water use of tropical cloud forest trees in response to experimental drought.

Tropical montane cloud forests (TMCFs) are expected to experience more frequent and prolonged droughts over the coming century, yet understanding of TCMF tree responses to moisture stress remains weak compared with the lowland tropics. We simulated a severe drought in a throughfall reduction experiment (TFR) for 2 years in a Peruvian TCMF and evaluated the physiological responses of several dominant species (Clusia flaviflora Engl., Weinmannia bangii (Rusby) Engl., Weinmannia crassifolia Ruiz & Pav. and Prunus integrifolia (C. Presl) Walp). Measurements were taken of (i) sap flow; (ii) diurnal cycles of stem shrinkage, stem moisture variation and water-use; and (iii) intrinsic water-use efficiency (iWUE) estimated from foliar δ13C. In W. bangii, we used dendrometers and volumetric water content (VWC) sensors to quantify daily cycles of stem water storage. In 2 years of sap flow (Js) data, we found a threshold response of water use to vapor pressure deficit vapor pressure deficit (VPD) > 1.07 kPa independent of treatment, though control trees used more soil water than the treatment trees. The daily decline in water use in the TFR trees was associated with a strong reduction in both morning and afternoon Js rates at a given VPD. Soil moisture also affected the hysteresis strength between Js and VPD. Reduced hysteresis under moisture stress implies that TMCFs are strongly dependent on shallow soil water. Additionally, we suggest that hysteresis can serve as a sensitive indicator of environmental constraints on plant function. Finally, 6 months into the experiment, the TFR treatment significantly increased iWUE in all study species. Our results highlight the conservative behavior of TMCF tree water use under severe soil drought and elucidate physiological thresholds related to VPD and its interaction with soil moisture. The observed strongly isohydric response likely incurs a cost to the carbon balance of the tree and reduces overall ecosystem carbon uptake.

Simulating forest productivity along a neotropical elevational transect: temperature variation and carbon use efficiency.

A better understanding of the mechanisms controlling the magnitude and sign of carbon components in tropical forest ecosystems is important for reliable estimation of this important regional component of the global carbon cycle. We used the JULES vegetation model to simulate all components of the carbon balance at six sites along an Andes-Amazon transect across Peru and Brazil and compared the results to published field measurements. In the upper montane zone the model predicted a lack of forest vegetation, indicating a need for better parameterization of the responses of cloud forest vegetation within the model. In the lower montane and lowland zones simulated ecosystem productivity and respiration were predicted with reasonable accuracy, although not always within the error bounds of the observations. Model-predicted carbon use efficiency in this transect surprisingly did not increase with elevation, but remained close to the 'temperate' value 0.5. Upper montane forests were predicted to allocate ~50% of carbon fixation to biomass maintenance and growth, despite available measurements showing that they only allocate ~33%. This may be explained by elevational changes in the balance between growth and maintenance respiration within the forest canopy, as controlled by both temperature- and pressure-mediated processes, which is not yet well represented in current vegetation models.

Impacts of insect frass and cadavers on soil surface litter decomposition along a tropical forest temperature gradient.

Insect herbivores play important roles in shaping many ecosystem processes, but how climate change will alter the effects of insect herbivory are poorly understood. To address this knowledge gap, we quantified for the first time how insect frass and cadavers affected leaf litter decomposition rates and nutrient release along a highly constrained 4.3°C mean annual temperature (MAT) gradient in a Hawaiian montane tropical wet forest. We constructed litterbags of standardized locally sourced leaf litter, with some amended with insect frass + cadavers to produce treatments designed to simulate ambient (Control = no amendment), moderate (Amended-Low = 2 × Control level), or severe (Amended-High = 11 × Control level) insect outbreak events. Multiple sets of these litterbags were deployed across the MAT gradient, with individual litterbags collected periodically over one year to assess how rising MAT altered the effects of insect deposits on litter decomposition rates and nitrogen (N) release. Increased MAT and insect inputs additively increased litter decomposition rates and N immobilization rates, with effects being stronger for Amended-High litterbags. However, the apparent temperature sensitivity (Q 10) of litter decomposition was not clearly affected by amendments. The effects of adding insect deposits in this study operated differently than the slower litter decomposition and greater N mobilization rates often observed in experiments which use chemical fertilizers (e.g., urea, ammonium nitrate). Further research is required to understand mechanistic differences between amendment types. Potential increases in outbreak-related herbivore deposits coupled with climate warming will accelerate litter decomposition and nutrient cycling rates with short-term consequences for nutrient cycling and carbon storage in tropical montane wet forests.

Integrating natural gradients, experiments, and statistical modeling in a distributed network experiment: An example from the WaRM Network.

A growing body of work examines the direct and indirect effects of climate change on ecosystems, typically by using manipulative experiments at a single site or performing meta-analyses across many independent experiments. However, results from single-site studies tend to have limited generality. Although meta-analytic approaches can help overcome this by exploring trends across sites, the inherent limitations in combining disparate datasets from independent approaches remain a major challenge. In this paper, we present a globally distributed experimental network that can be used to disentangle the direct and indirect effects of climate change. We discuss how natural gradients, experimental approaches, and statistical techniques can be combined to best inform predictions about responses to climate change, and we present a globally distributed experiment that utilizes natural environmental gradients to better understand long-term community and ecosystem responses to environmental change. The warming and (species) removal in mountains (WaRM) network employs experimental warming and plant species removals at high- and low-elevation sites in a factorial design to examine the combined and relative effects of climatic warming and the loss of dominant species on community structure and ecosystem function, both above- and belowground. The experimental design of the network allows for increasingly common statistical approaches to further elucidate the direct and indirect effects of warming. We argue that combining ecological observations and experiments along gradients is a powerful approach to make stronger predictions of how ecosystems will function in a warming world as species are lost, or gained, in local communities.