Abstract
The response of root respiration to warmer soil can affect ecosystem carbon (C) allocation and the strength of positive feedbacks between climatic warming and soil CO2 efflux. This study sought to determine whether fine-root (<1 mm) respiration in a sugar maple (Acer saccharum Marsh.)-dominated northern hardwood forest would adjust to experimentally warmed soil, reducing C return to the atmosphere at the ecosystem scale to levels lower than that would be expected using an exponential temperature response function. Infrared heating lamps were used to warm the soil (+4 to +5 °C) in a mature sugar maple forest in a fully factorial design, including water additions used to offset the effects of warming-induced dry soil. Fine-root-specific respiration rates, root biomass, root nitrogen (N) concentration, soil temperature and soil moisture were measured from 2009 to 2011, with experimental treatments conducted from late 2010 to 2011. Partial acclimation of fine-root respiration to soil warming occurred, with soil moisture deficit further constraining specific respiration rates in heated plots. Fine-root biomass and N concentration remained unchanged. Over the 2011 growing season, ecosystem root respiration was not significantly greater in warmed soil. This result would not be predicted by models that allow respiration to increase exponentially with temperature and do not directly reduce root respiration in drier soil.
Introduction
The fine roots (<1 mm) of woody perennials can be very active contributors to ecosystem autotrophic respiration. In northern hardwood forests, very fine roots (<0.5 mm) in surface soil (0–10 cm soil depth) have been shown to contribute 53% of ecosystem root respiration and those to a depth of 50 cm to contribute 69% of ecosystem root respiration (Burton et al. 2012). Plant tissue respiration has been found to increase exponentially in response to immediate increases in temperature (Ryan et al. 1996, Tjoelker et al. 2001, Piao et al. 2010) with a Q10 often between 1.8 and 3.0 for roots (Burton et al. 2002, Piao et al. 2010). If this response to temperature held true for a long-term climatic warming, exponentially more photosynthate might be returned to the atmosphere as CO2 at higher temperatures, with less carbon (C) left for net primary production. This could accelerate CO2 build up in the atmosphere, as some C that could have been sequestered in plant biomass was released through autotrophic respiration. A positive feedback loop could occur where increased temperatures would cause more CO2 to return to the atmosphere, causing even higher global temperatures (Woodwell and Mackenzie 1995).
This positive feedback loop would be lessened or eliminated if the plants acclimated to these warmer temperatures by reducing metabolic activity of existing tissues (Atkin and Tjoelker 2003) or by creating new less metabolically active tissues when ephemeral components, such as leaves and fine roots, are replaced (Loveys et al. 2003, Ow et al. 2008). For roots, respiratory activity should be related to the energy required to produce new roots, take up and transport ions, and repair tissues. Since such needs, with the possible exception of tissue repair, are unlikely to increase exponentially with temperature at the ecosystem level, an exponential increase in ecosystem root respiration also would not be expected. This is in contrast to the temperature responses for respiration used in many vegetation models that serve as components of earth systems (Smith and Dukes 2013).
The presence and degree of acclimation can vary greatly among plant tissues and species. In foliage, acclimation has ranged from homeostasis (complete acclimation) to no acclimation (Larigauderie and Körner 1995, Atkin et al. 2000a, 2000b, Loveys et al. 2003, Cooper 2004). For roots, fewer studies exist, but again a wide range of responses, from little acclimation to levels approaching homeostasis, have been found in response to altered growth temperatures (Bryla et al. 1997, 2001, Tjoelker et al. 1999, Atkinson et al. 2007, Burton et al. 2008, Rachmilevitch et al. 2008). Many of these studies involved seedlings in growth chambers or greenhouses. Data for roots of larger, field grown trees are more limited and, in some cases, responses depended on other environmental conditions. For example, Bryla et al. (1997) found growth in different moisture regimes affected temperature acclimation of Citrus volkameriana V. Ten & Pasq. root respiration, with acclimation occurring in wet soils but not in trees growing in dry soils. In a separate study, acclimation of root respiration in citrus was also found to occur only for temperature >23 °C (Bryla et al. 2001). Little to no acclimation occurred in response to seasonal temperature variations for root respiration in Pinus resinosa Ait. plantations and Acer saccharum Marsh. forests (Burton and Pregitzer 2003). They speculated that natural fluctuations of up to 5–10 °C in field soil temperature, often within a few days, may have limited the degree of acclimation occurring. They also argued that down-regulation of root activity during warm, moist periods could be counter-productive, as it would reduce nutrient uptake capacity at times of enhanced nutrient availability associated with mineralization of organic matter.
In response to chronic elevated soil temperatures, a variety of mechanisms could allow ecosystem root-system respiration to adjust. At the tissue/cellular level, changes in enzymatic capacity, substrate limitation and adenylate control all could cause temperature acclimation (Atkin and Tjoelker 2003). Additionally, at the ecosystem level, a reduction in fine-root biomass has recently been shown to be a key factor in constraining ecosystem fine-root respiration in hardwood forests subjected to long-term soil warming (Melillo et al. 2011), despite the lack of temperature acclimation at the tissue scale (i.e., no down-regulation of respiration per unit mass).
The objective of this study was to determine whether fine-root (<1 mm) respiration in a sugar maple-dominated northern hardwood forest would adjust to experimentally warmed soil temperatures in the short term (days to weeks) or longer term (months), reducing C return to the atmosphere at the ecosystem scale to levels less than that would be predicted from an exponential temperature response function for respiration. The study was conducted in the Upper Peninsula of Michigan, where growing season temperature is predicted to increase by at least 4 °C during the next century, with little change in mean annual precipitation and possible slight decreases in growing season precipitation (Christensen et al. 2007). The soil was experimentally warmed (+4 to +5 °C) with the use of infrared heating lamps in a factorial combination with water addition. The water addition was intended to allow the effects of warmer soil on fine-root respiration to be separated from effects created by co-occurring drier soil conditions. The factorial combination of soil warming and moisture addition is the only field experiment of its kind in a mature temperate hardwood forest.
We hypothesized that short-term (days to weeks) acclimation to soil warming would not occur, leading to temporarily enhanced CO2 efflux from fine-root respiration as an initial response to soil warming. We also hypothesized that longer term (months to years) adjustment of fine-root respiration at the ecosystem level would occur. This was expected as an exponential increase in fine-root respiration would likely cause respiratory activity to be greater than that needed to facilitate the work performed by the root system for nutrient uptake, assimilation and transport, as well as cell growth and maintenance. Two possible mechanisms for such a response at the ecosystem level were assessed in this study: (i) a change over time in the enzymatic capacity of fine roots produced in heated soil, with a shift to fine roots with lower enzymatic capacity, indicated by lower root nitrogen (N) concentrations; and (ii) a reduction in fine-root biomass in warmer soils.
Materials and methods
Site description
The study was conducted at Michigan Technological University's Ford Forestry Center (FFC) in Baraga County, Michigan (46°38′26.17″N 88°29′00.94″W, 400 m elevation) during the growing seasons of 2009 to 2011. Pretreatment measurements were made in 2009 and early 2010, and experimental measurements in late 2010 and all of 2011. The mean annual temperature in this region is 4.9 °C, with a growing season (May to September) average temperature of 15 °C. The area receives on average 879 mm of precipitation annually, with 401 mm of precipitation falling during the growing season (Burton et al. 2012).
Sugar maple dominates the overstory (trees >5.0 cm diameter), contributing 89.3% (21.7 m2 ha−1) of the overstory basal area. American elm (Ulmus americana L.), eastern hemlock (Tsuga canadensis (L.) Carr.), ironwood (Ostrya virginiana (Mill.) K. Koch) and yellow birch (Betula alleghaniensis Britton) comprise the remainder. Dominant trees in the overstory are often 100 years or more in age. The understory consists of young sprouts and seedlings of overstory trees with the addition of black cherry (Prunus serotina Ehrh.), which rarely competes successfully in the overstory at this location. The herbaceous layer consists of American fly honeysuckle (Lonicera canadensis Bartram ex Marsh.), common lady fern (Athyrium filix-femina (L.) Roth), spinulose shield fern (Dryopteris carthusiana (Vill.) H.P. Fuchs), wild leek (Allium burdickii (Hanes) A.G. Jones), Dutchman's breeches (Dicentra cucullaria (L.) Bernh.), trillium (Trillium spp. L.) and yellow trout lily (Erythronium americanum Ker.).
The soil at the site is classified as a Kallio cobbly silt loam (coarse-loamy, mixed, superactive, frigid Oxyaquic Fragiorthods), which consists of a cobbly silt loam and silt loam to a depth of 40 cm below soil surface. A fragipan occurs at ∼40 cm, and soil textures below this depth include sandy loams and gravelly loams to depths of 150 cm.
Experimental treatments
Twelve 10 × 10 m research plots were established in 2009. These were divided into three groups of four plots based on geographic separation, with each of four treatments randomly assigned to one plot in each group. Experimental treatments included a control (no treatment), soil warming (+4 to +5 °C above ambient) by infrared lamps, water addition (ambient + 30% of average growing season ambient) and soil warming plus water addition. The water additions are intended to offset the increased evaporative water loss due to warming. Wooden boardwalks were installed throughout the plots to minimize disturbance and soil compaction during installation of equipment and sample collection.
Sixteen infrared heating lamps (model MRM1215 heaters, Kalglo Electronics Co., Bethlehem, PA, USA) were suspended 1.5 m above the soil in four rows of four lamps, all at 2.5 m spacing. The lamps were operated 24 h per day from 17 September to 16 November in 2010 and from 22 April to 11 November in 2011. Lamp output was manually adjusted as needed to maintain targeted soil temperatures. Soil temperature measurements on site and field observations after snowfall events indicated the actual heated area slightly exceeded the desired 10 × 10 m area, with soil temperature gradually declining for up to 2 m beyond the plot boundary.
Soil temperature and soil moisture were monitored with data loggers recording at 30-min intervals. Volumetric soil moisture and soil temperature were recorded at 2, 5 and 10 cm below the soil surface under heater rows and at 2 and 5 cm depths midway between heater rows in locations approximately midway from plot center to plot edge (Em50 data loggers with 5TM temperature/moisture probes, Decagon Devices, Inc., Pullman, Washington, DC, USA). At plot center, soil temperature at depths of 1, 5 and 15 cm and air temperature at 1 m were monitored (Hobo U12 4-external channel outdoor/ industrial data logger with TMC6-HA probes, Onset Computer Corporation, Bourne, MA, USA). Additionally, on plots receiving warming, soil temperature was monitored at 2 and 5 cm below the soil surface directly under and halfway between heater rows near the plot edge (Hobo U12 4-external channel outdoor/industrial data logger with TMC6-HA probes, Onset Computer Corporation).
Precipitation used for the water additions was captured in six 1900-l tanks, using gutter systems on the rooftop of the dormitory building at the Ford Forestry Center. Filled tanks were transported to areas on woods roads near (within 50 m) the experimental plots using an all-terrain forklift, and water was distributed to the plots by pumping from the tanks to four sprinkler heads (5000 series rotor, Rainbird Corporation, Tucson, AZ, USA) per plot. Typically 0.95 cm of rainfall depth equivalent was added once per week. When possible, water additions were made as supplements to natural rain events, to allow for natural wetting and drying cycles on watered plots. The total rainfall equivalent depth added in 2011 was 15.8 cm. Water additions were used to offset soil drying associated with increased soil temperatures for the heat treatments. The needed growing season water addition was based on estimates of additional evapotranspiration that would occur if the mean climate at the site warmed by 5 °C and is equivalent to ∼30% of the long-term mean growing season precipitation.
Fine-root respiration and biomass
Fine-root (<1 mm diameter)-specific respiration (nmol CO2 g−1 dry weight s−1) was measured at 2- to 4-week intervals from April to October in 2009 to 2011 using open-system infrared gas analyzers (IRGAs, CIRAS-1 and CIRAS-2 portable gas analyzers, PP Systems, Haverhill, MA, USA). Measurements were made on 13 dates during the pretreatment period (27 May 2009 to 30 June 2010), four dates during a post-installation disturbance period (23 July to 13 September 2010) and 12 dates after initiation of treatments (20 September 2010 to 11 September 2011). On each sample date, specific fine-root respiration rates were determined at both ambient soil temperature for the treatment and at a constant reference temperature of 18 °C. Measurements at the reference temperature were used to assess changes in respiratory capacity over time and across treatments, and have been found to be a reliable test of acclimation to experimental warming (Atkin et al. 2000a). We chose 18 °C as the common reference temperature to assess acclimation, as it approximates the soil temperature during the warmest period of the growing season for the control plots.
Excised fine-root samples for respiration measurement were obtained from each plot using two 5-cm diameter by 10-cm deep soil cores, taken just below the loose surface litter (Oi) layer. Cores contained 1–2 cm of forest floor (Oe and Oa horizons) and 8–9 cm of surface mineral soil (primarily A horizon, with small amounts of E horizon in some instances). These samples were taken from the inner 5 × 5 m portion of the experimental plots, to ensure that they contained roots from trees for which a majority of the root system was warmed. Roots were hand sorted from the cores for a 10- to 15-min period and brushed free of soil and organic matter. When possible, intact fine-root branching networks, consisting largely of first- and second-order roots, were detached intact from larger diameter roots, to minimize root damage. Live roots were distinguished from dead roots by white, tan or brown coloration and a smooth appearance. Dead roots were brittle, had rough edges and dark brown or black coloration. Visual observations of root morphology during sorting suggested that the proportion of sample mass comprised of sugar maple roots was similar to the proportion of sugar maple basal area at the site (∼90%). In this study, we focused on fine roots in the top 10 cm of soil, as previous work in similar northern hardwood forests has shown that surface fine roots contribute 60% of fine-root biomass to a depth of 50 cm (Burton et al. 2012). The surface fine roots are also by far the most active part of the root system in these forests, contributing 75% of fine-root respiration and 53% of total root-system respiration in sugar maple-dominated northern hardwood forests (Burton et al. 2012).
Approximately 2 g (fresh weight) of fine roots were placed in a respiration cuvette attached to an IRGA (Burton and Pregitzer 2003, Burton et al. 2012). Respiration rates were recorded after allowing 15 min for readings to stabilize. The cuvette bases were placed in water baths to maintain the respiration samples at the desired temperatures. Each individual sample was measured at both the ambient and reference temperatures, with the order of temperatures alternated among samples to avoid any potential biases associated with time since sample collection. Respiration was analyzed at a CO2 concentration of 1000 μl l−1, which approximates the soil CO2 concentration typically found near the soil surface of sugar maple-dominated northern hardwood forests (Burton et al. 1997). Measurements were typically completed within 30–45 min of taking the soil cores. After the measurement, samples were placed on ice until they could be returned to the laboratory, where they were frozen pending further analysis. The samples were later thawed, cleaned with deionized water to remove any soil and organic matter not removed in the field (typically <2% of sample weight) and dried at 65 °C for 48 h. The samples were then ground to a fine powder (8000 M Mixer/Mill, Spex SamplePrep LLC, Metuchen, NJ, USA) and analyzed for N concentration with an elemental analyzer (Carlo Erba NA 1500 NC, CE Elantech, Lakewood, NJ, USA).
Fine-root biomass was assessed in late August 2010 and early September 2011 using eight soil cores, 5 cm in diameter to a depth of 10 cm, per plot. The cores were sieved (2 mm mesh) to remove a majority of roots, with additional roots obtained by hand sorting the core material. Live roots were distinguished by color (white to brown) and physical integrity and then separated into diameter classes, cleaned with deionized water, dried for 48 h and weighed. This effort provided a much more complete estimate of fine-root biomass than the brief field sorting of soil cores during respiration measurements. Ecosystem-level fine-root respiration for dates within the 2011 growing season was calculated as a product of measured specific fine-root respiration rates and 2011 fine-root biomass. Total growing season (May–September) fine-root respiration was estimated by assuming that the value for each measurement date was applicable to all days within a period extending halfway to both the prior and next sampling dates.
Since N availability can have significant impacts on root-system activity, net N mineralization rates were determined using the buried bag technique (Eno 1960), with three incubations per plot initiated every 4–5 weeks from 2 May to 26 October 2011. At the beginning of each measurement interval, three randomly located pairs of soil cores (5 cm diameter by 10 cm in depth) were collected from each plot. One of each pair was taken back to the laboratory, and a 20-g subsample was extracted with 40 ml of 2 M KCl for determination of initial soil extractable NO3− and NH4+ contents. The other core from each pair was placed in a polyethylene bag, replaced into its original hole and incubated in situ until the beginning of the next measurement period. Subsamples from these cores were then extracted with 2 M KCl to assess final soil NO3− and NH4+ contents. For each measurement period, net N mineralization was calculated as the difference in NO3− -N between final and initial cores of a pair. Values from the three sample locations per plot were averaged to determine a plot-level value for each measurement interval, and these were summed across the entire measurement period and divided by 177 days to determine an average net N mineralization rate per day.
Statistical analysis
Non-linear regression was used to develop temperature response curves for specific fine-root respiration at ambient soil temperature using data from the pretreatment period for all plots plus data from control plots during the treatment period. These same data were used to assess acclimation to seasonal variations in soil temperature by regressing specific fine-root respiration at the reference temperature of 18 °C against the applicable ambient soil temperatures for the sample dates. A strong negative correlation would indicate that, as seasonal temperatures increased, fine roots acclimated to warmer temperatures by down-regulating specific respiration rate. Repeated-measures analysis of variance (ANOVA) was used to test the effects of soil warming and water additions on fine-root respiration rates across time. Separate analyses were performed for the pretreatment, installation disturbance and treatment periods and for ambient and reference temperatures. The pretreatment and disturbance analyses used date (repeated measure) and future treatment as factors in the ANOVA. The treatment period analysis used a two-factor (heat and water) repeated-measures (date) ANOVA. A heat × date interaction term for the 18 °C reference temperature in this analysis was further investigated using analysis of covariance. This analysis sought to determine whether a downward adjustment in fine-root metabolic capacity for heated treatments still existed after accounting for the influence of the covariate, soil moisture content.
Results
Pretreatment analysis
There were no pre-existing differences in specific fine-root respiration rates among the groups of plots assigned to each treatment at ambient temperature (P = 0.69) or the 18 °C reference temperature (P = 0.70) (see Supplementary Figure 1a available as Supplementary Data at Tree Physiology Online). After the heating lamps and sprinklers had been installed, but before treatments were initiated, respiration was measured on four additional sample dates. Again, no differences were found among the groups of plots assigned to each treatment (P = 0.94 for both ambient and reference temperature), indicating fine-root respiration rates had not been altered by any disturbance from infrastructure installation (Supplementary Figure 1b available as Supplementary Data at Tree Physiology Online).
Temperature response curve for specific fine-root respiration at ambient soil temperature from May 2009 to August 2011. Data are from all plots prior to initiation of treatments in September 2010 and from the control plots only, thereafter.
Using data from all treatments prior to initiation of soil warming and from the control plots after treatments were initiated, specific fine-root respiration was shown to exhibit an exponential increase to seasonal changes in soil temperature with a Q10 of 2.7 (Figure 1). Specific fine-root respiration at the 18 °C reference temperature, used as an indicator of metabolic capacity, was not correlated with ambient soil temperature for the unheated control plots during the study (Supplementary Figure 2 available as Supplementary Data at Tree Physiology Online).
Effects of experimental treatments on soil temperature (a) and soil moisture (b) during 2011. Values are the mean of all sensors for a given treatment. Temperature enhancement (+4.9 °C) was similar for all depths (2, 5 and 10 cm) and locations (beneath and between heater lines in plot center and near plot edge).
Responses to soil warming and moisture addition
The warming lamps successfully raised soil temperatures throughout the heated plots by the intended 4–5 °C (Figure 2a), with an average soil temperature increase of 4.9 °C across the treatment periods of 2010 and 2011. We also successfully maintained volumetric soil moisture for the heat + water treatments at levels close to those observed in the control plots (Figure 2b). In 2011, soils became quite dry in all treatments from mid-July to early September (Figure 2b), in association with a pronounced regional drought.
Repeated-measures ANOVA indicated that the effect of soil warming and moisture additions on specific fine-root respiration at ambient soil temperature varied among sampling dates (Table 1), leading to a significant heat × date interaction (P = 0.018) and a nearly significant water × date interaction (P = 0.062). When soil moisture was sufficient, soil warming promoted elevated fine-root respiration rates (Figure 3a, dates through June 2011), as indicated by a positive effect of the heat treatment (P = 0.026), with no heat × date interaction, when the analysis was run only for the sample dates through June 2011, when soils were moist. However, beginning in July 2011, when soils began to dry, severely so in the heat-only treatment, specific root respiration at ambient temperature was no longer significantly greater in heated plots (P = 0.177 for the heat effect in repeated-measures ANOVA for July–September, 2011), despite their warmer temperature, and often was lower (Figure 3a).
Repeated-measures ANOVA for the effects of soil warming (heat) and moisture addition (water) on specific fine-root respiration rate (nmol g−1 s−1) at ambient soil temperature and at a reference temperature of 18 °C. df, degrees of freedom.
| Source | df | Ambient soil temperature | 18 °C reference | ||||
|---|---|---|---|---|---|---|---|
| Mean square | F-ratio | P | Mean square | F-ratio | P | ||
| Between subjects | |||||||
| Heat | 1 | 2.50 | 1.16 | 0.314 | 47.47 | 18.40 | 0.003 |
| Water | 1 | 6.48 | 3.00 | 0.121 | 1.54 | 1.60 | 0.462 |
| Heat × water | 1 | 0.76 | 0.35 | 0.569 | 3.66 | 1.42 | 0.268 |
| Error | 8 | 2.16 | 2.58 | ||||
| Within subjects | |||||||
| Sample date Sample date | 11 | 9.04 | 11.44 | <0.001 | 24.59 | 15.41 | <0.001 |
| Heat × date | 11 | 1.78 | 2.26 | 0.018 | 1.61 | 1.01 | 0.448 |
| Water × date | 11 | 1.44 | 1.82 | 0.062 | 1.81 | 1.13 | 0.345 |
| Heat × water × date | 11 | 0.34 | 0.43 | 0.938 | 0.50 | 0.31 | 0.981 |
| Error | 88 | 0.79 | 1.60 | ||||
| Source | df | Ambient soil temperature | 18 °C reference | ||||
|---|---|---|---|---|---|---|---|
| Mean square | F-ratio | P | Mean square | F-ratio | P | ||
| Between subjects | |||||||
| Heat | 1 | 2.50 | 1.16 | 0.314 | 47.47 | 18.40 | 0.003 |
| Water | 1 | 6.48 | 3.00 | 0.121 | 1.54 | 1.60 | 0.462 |
| Heat × water | 1 | 0.76 | 0.35 | 0.569 | 3.66 | 1.42 | 0.268 |
| Error | 8 | 2.16 | 2.58 | ||||
| Within subjects | |||||||
| Sample date Sample date | 11 | 9.04 | 11.44 | <0.001 | 24.59 | 15.41 | <0.001 |
| Heat × date | 11 | 1.78 | 2.26 | 0.018 | 1.61 | 1.01 | 0.448 |
| Water × date | 11 | 1.44 | 1.82 | 0.062 | 1.81 | 1.13 | 0.345 |
| Heat × water × date | 11 | 0.34 | 0.43 | 0.938 | 0.50 | 0.31 | 0.981 |
| Error | 88 | 0.79 | 1.60 | ||||
Repeated-measures ANOVA for the effects of soil warming (heat) and moisture addition (water) on specific fine-root respiration rate (nmol g−1 s−1) at ambient soil temperature and at a reference temperature of 18 °C. df, degrees of freedom.
| Source | df | Ambient soil temperature | 18 °C reference | ||||
|---|---|---|---|---|---|---|---|
| Mean square | F-ratio | P | Mean square | F-ratio | P | ||
| Between subjects | |||||||
| Heat | 1 | 2.50 | 1.16 | 0.314 | 47.47 | 18.40 | 0.003 |
| Water | 1 | 6.48 | 3.00 | 0.121 | 1.54 | 1.60 | 0.462 |
| Heat × water | 1 | 0.76 | 0.35 | 0.569 | 3.66 | 1.42 | 0.268 |
| Error | 8 | 2.16 | 2.58 | ||||
| Within subjects | |||||||
| Sample date Sample date | 11 | 9.04 | 11.44 | <0.001 | 24.59 | 15.41 | <0.001 |
| Heat × date | 11 | 1.78 | 2.26 | 0.018 | 1.61 | 1.01 | 0.448 |
| Water × date | 11 | 1.44 | 1.82 | 0.062 | 1.81 | 1.13 | 0.345 |
| Heat × water × date | 11 | 0.34 | 0.43 | 0.938 | 0.50 | 0.31 | 0.981 |
| Error | 88 | 0.79 | 1.60 | ||||
| Source | df | Ambient soil temperature | 18 °C reference | ||||
|---|---|---|---|---|---|---|---|
| Mean square | F-ratio | P | Mean square | F-ratio | P | ||
| Between subjects | |||||||
| Heat | 1 | 2.50 | 1.16 | 0.314 | 47.47 | 18.40 | 0.003 |
| Water | 1 | 6.48 | 3.00 | 0.121 | 1.54 | 1.60 | 0.462 |
| Heat × water | 1 | 0.76 | 0.35 | 0.569 | 3.66 | 1.42 | 0.268 |
| Error | 8 | 2.16 | 2.58 | ||||
| Within subjects | |||||||
| Sample date Sample date | 11 | 9.04 | 11.44 | <0.001 | 24.59 | 15.41 | <0.001 |
| Heat × date | 11 | 1.78 | 2.26 | 0.018 | 1.61 | 1.01 | 0.448 |
| Water × date | 11 | 1.44 | 1.82 | 0.062 | 1.81 | 1.13 | 0.345 |
| Heat × water × date | 11 | 0.34 | 0.43 | 0.938 | 0.50 | 0.31 | 0.981 |
| Error | 88 | 0.79 | 1.60 | ||||
Specific fine-root respiration at ambient temperature (a) and the 18 °C reference temperature (b) from 20 September 2010 to 12 September 2011. Differences among sample dates in average fine-root respiration reflect differences in ambient soil temperature (Figure 2), as well as the effects of dry soils from 15 July to 12 September 2011. Soil temperatures in the heat and heat + water treatments were ∼5 °C warmer than those in the control and water-only treatments. Error bars are one standard error of the mean.
Fine-root-specific respiration rates at the 18 °C reference temperature were always lower for the heat treatment than those for the control (P = 0.003), with the effect more pronounced when soils were dry (Figure 3b, P = 0.067 through June 2011 vs P < 0.001 for July–September 2011). Beginning in mid-July 2011, drier soil reduced fine-root metabolic capacity for fine roots of all treatments, but effects were much greater for the treatments with soil warming (Figure 3b). Analysis of covariance was used to compare the relationship between soil moisture content and fine-root respiration at the 18 °C reference temperature among treatments. Moisture × treatment interactions were absent (P = 0.331), indicating that regression slopes were similar among treatments (Figure 4). A significant heat effect existed (P = 0.005), indicating that regression intercepts for the heat and heat + water treatments were lower than those for the control and water treatments (Figure 4).
Relationships between specific fine-root respiration at the 18 °C reference temperature and volumetric soil moisture by treatment. Regression relationships for all treatments are significant (P ≤ 0.006). Those for the heat and heat + water treatments have significantly lower intercepts (P = 0.005) than those for the control and water treatments. Data points shown are sample date means for all three plots receiving a treatment.
Net N mineralization during the growing season was ∼60% greater (P = 0.040) for treatments receiving soil warming than that for those which did not (Table 2). Surface fine-root biomass did not differ significantly among treatments for 2010 (P = 0.88) or 2011 (P = 0.93), and averaged 373 g m−2 across all treatments in 2011 (Table 2). Fine-root N concentration was also unaffected by the experimental treatments (Table 2). When surface fine-root biomass from each plot is used in conjunction with fine-root-specific respiration rates for the plot to estimate the soil CO2 efflux from fine roots during the 2011 growing season, average values were 2, 24 and 26% greater than the control for the heat, water and heat + water treatments, respectively (Figure 5), but treatment effects were not statistically significant.
Surface fine-root (<1 mm) biomass in 2011, fine-root N concentration and soil net N mineralization for the four experimental treatments. Standard error of the mean is presented in parentheses. Fine-root biomass and N concentration were not affected by the treatments. Net N mineralization was significantly greater on heated plots (P = 0.040).
| Control | Heat | Water | Heat + water | |
|---|---|---|---|---|
| Surface fine-root biomass (g m−2)1 | 344 (15) | 370 (63) | 391 (58) | 389 (89) |
| Fine-root N, pretreatment (g N kg − 1) | 11.7 (0.3) | 11.5 (0.3) | 11.5 (0.3) | 11.5 (0.3) |
| Fine-root N, treatment period (g N kg − 1) | 11.0 (0.2) | 11.1 (0.2) | 11.3 (0.2) | 10.9 (0.2) |
| N mineralization (mg N g soil − 1 day − 1) | 11.7 (0.6) | 19.3 (4.6) | 10.4 (2.4) | 16.6 (2.3) |
| Control | Heat | Water | Heat + water | |
|---|---|---|---|---|
| Surface fine-root biomass (g m−2)1 | 344 (15) | 370 (63) | 391 (58) | 389 (89) |
| Fine-root N, pretreatment (g N kg − 1) | 11.7 (0.3) | 11.5 (0.3) | 11.5 (0.3) | 11.5 (0.3) |
| Fine-root N, treatment period (g N kg − 1) | 11.0 (0.2) | 11.1 (0.2) | 11.3 (0.2) | 10.9 (0.2) |
| N mineralization (mg N g soil − 1 day − 1) | 11.7 (0.6) | 19.3 (4.6) | 10.4 (2.4) | 16.6 (2.3) |
1Data are for fine roots from the top 10 cm of forest floor and mineral soil.
Surface fine-root (<1 mm) biomass in 2011, fine-root N concentration and soil net N mineralization for the four experimental treatments. Standard error of the mean is presented in parentheses. Fine-root biomass and N concentration were not affected by the treatments. Net N mineralization was significantly greater on heated plots (P = 0.040).
| Control | Heat | Water | Heat + water | |
|---|---|---|---|---|
| Surface fine-root biomass (g m−2)1 | 344 (15) | 370 (63) | 391 (58) | 389 (89) |
| Fine-root N, pretreatment (g N kg − 1) | 11.7 (0.3) | 11.5 (0.3) | 11.5 (0.3) | 11.5 (0.3) |
| Fine-root N, treatment period (g N kg − 1) | 11.0 (0.2) | 11.1 (0.2) | 11.3 (0.2) | 10.9 (0.2) |
| N mineralization (mg N g soil − 1 day − 1) | 11.7 (0.6) | 19.3 (4.6) | 10.4 (2.4) | 16.6 (2.3) |
| Control | Heat | Water | Heat + water | |
|---|---|---|---|---|
| Surface fine-root biomass (g m−2)1 | 344 (15) | 370 (63) | 391 (58) | 389 (89) |
| Fine-root N, pretreatment (g N kg − 1) | 11.7 (0.3) | 11.5 (0.3) | 11.5 (0.3) | 11.5 (0.3) |
| Fine-root N, treatment period (g N kg − 1) | 11.0 (0.2) | 11.1 (0.2) | 11.3 (0.2) | 10.9 (0.2) |
| N mineralization (mg N g soil − 1 day − 1) | 11.7 (0.6) | 19.3 (4.6) | 10.4 (2.4) | 16.6 (2.3) |
1Data are for fine roots from the top 10 cm of forest floor and mineral soil.
Estimated ecosystem fine-root respiration (g C m−2) for the four treatments during the 2011 growing season (May–September). Error bars are one standard error of the mean for three plots per treatment. Ecosystem fine-root respiration was not affected by soil warming (P = 0.905), water treatments (P = 0.160) or their interaction (P = 0.977).
Discussion
We had hypothesized that immediate acclimation of fine-root respiration would not occur in response to soil warming, causing fine-root respiration to be greater in warmed soil. This hypothesis was based in part on the lack of acclimation in fine-root respiration in response to seasonal variations in soil temperature (Burton and Pregitzer 2003), which also occurred for the control plots in this study during the pretreatment measurement period (Supplementary Figure 1 available as Supplementary Data at Tree Physiology Online). In these tests for roots from non-warmed soil, fine-root respiration rates at a given temperature and respiratory Q10 were the same, whether measured across the growing season at different ambient temperatures or measured using a temperature series analyzed on a single day (Burton and Pregitzer 2003). As a result, it was expected that the respiration rate in warmed soils would increase in accordance with the same exponential temperature response during the first weeks and months following initiation of soil warming. This hypothesis was only partially supported, as fine-root respiration for the heat and heat + water treatments was markedly greater than that for the control during September to November in 2010, immediately after initiation of warming, and again in May and June 2011. However, the increase in respiration rates (13–29%) was far less than the 49–64% increase that would be predicted for the 4–5 °C increase in soil temperature using the observed Q10 for the study location (Q10 = 2.7, Figure 1). Clearly, soon after initiation of soil warming, some degree of acclimation had occurred. Lower respiration rates at the 18 °C reference temperature for the heat and heat + water treatments are another indication of this partial acclimation.
Dry soil conditions also had a major impact on fine-root respiration, causing much lower rates at the 18 °C reference temperature for all treatments in the latter half of the 2011 growing season (Figure 3b) when soils became dry. The effect was especially pronounced for the heat-only treatment, in which soil moisture deficits were most severe (Figure 2b), leading to respiration rates at ambient soil temperatures that were lower than those in the control plots (Figure 3b), despite soil temperatures that were 5 °C warmer. The effects of drier soils can account for some of the apparent acclimation for the heat treatment, but when soil moisture is accounted for, temperature acclimation of fine-root respiration is still apparent for the heat and heat + water treatments. Significantly lower intercepts for heat and heat + water regressions in Figure 4 indicate that regardless of soil moisture content, fine-root metabolic capacity, as indicated by respiration at the 18 °C reference temperature, was lower for treatments involving warming.
At the ecosystem level, surface fine-root respiration was not significantly greater than the control for any of the treatments during the 2011 growing season. In the heat-only treatment, the effects of dry soil in combination with acclimation caused the soil CO2 efflux from fine-root respiration to be nearly identical to the control treatment (Figure 5). Despite this, ecosystem C balance for the heat treatment may still have been altered, as the very dry soil conditions may have led to reductions in photosynthetic C assimilation due to stomatal closure in response to decreased plant water availability. In the heat + water treatment, soil moisture was maintained at levels near the control (Figure 2b), with growing season volumetric soil moisture averaging 0.213 and 0.208 for the control and water + heat treatments, respectively. Under these conditions, moisture limitation of photosynthesis by soil moisture availability should be similar for the control and heat + water treatments. If photosynthesis for the treatments was similar, the 26% greater ecosystem fine-root respiration for the heat + water treatment, although not statistically greater than the control, would have the potential to adversely impact the amount of C remaining for net primary productivity. Still, due to acclimation, any such effects were far less than that would have potentially occurred in the absence of adjustment in respiration rates.
Several mechanisms exist that could explain our observed partial temperature acclimation for fine-root respiration, with substrate limitation and adenylate control felt to be the more likely causes for acclimation at higher temperatures (Atkin and Tjoelker 2003). Substrate limitation could occur if the supply of carbohydrates was insufficient to support the respiratory needs of the root tissues. We have not observed seasonal temperature acclimation in the control plots, suggesting that considerable carbohydrate supply is available to support high levels of root activity during occasional warm periods. However, sustained warmer temperatures create a greater possibility that the tree's overall allocation of carbohydrate to root systems could be exceeded by potential respiratory demand. Adenylate control in fine roots in heated soil could result from the ATP requirement for growth and nutrient uptake, assimilation and transport being less than the respiratory pathway's ability to generate ATP at the warmer temperature. In such a case, incomplete use of available ATP to perform work would result in reduced regeneration of ADP, leading to low ADP concentrations and reduced respiration (Atkin and Tjoelker 2003). Increases in N availability in the heat and heat + water treatments (Table 2) suggest that the potential exists for significantly enhanced N uptake, assimilation and transport and thus ATP requirement. However, if the greater soil N availability was in excess of tree requirements, large increases in N uptake and assimilation may not have occurred, potentially leading to adenylate control.
We have not warmed the canopy in this experiment, and as a result several possible effects of a warmer climate on ecosystem C balance could not occur. These include both potential positive effects, such as an increase in growing season length, and negative effects, including possible increases in foliar respiration or greater occurrence of stomatal closure induced by vapor pressure deficit. Such factors could alter the amount of C available for allocation to the root system, but we feel our findings suggest that C used by the root system in warmer soil will remain in line with the work required for providing needed water and nutrients.
We had hypothesized that a longer-term response to warming would be the construction of new fine roots with lower enzyme concentrations, as indicated by lower tissue N concentration. Under field conditions it was hypothesized that greater specific enzyme activity in fine roots in warmer soil would counteract the effects of lower enzyme concentrations, enabling sufficient metabolic activity to still occur, but without excessive respiratory C losses. Construction of tissues with lower N concentration in warmer environments (Tjoelker et al. 1999) and adjustment of tissue N in response to temperature (Lee et al. 2005, Tjoelker et al. 2008) have both been observed for foliage. Thus far, we have seen no evidence of such adjustments in enzymatic capacity occurring for fine roots in our experiment, as fine-root N concentrations were not affected by the treatments (Table 2). However, it should be noted that adjustment of tissue N as a means of temperature acclimation is thought to be more likely for new tissue produced in the altered environment (Loveys et al. 2003). Given the fairly low turnover rates for sugar maple fine roots (0.5–0.7 year−1, Burton et al. 2000), many of the fine roots we sampled early in the experimental warming may have developed prior to initiation of soil warming. By the fall of 2011, however, well over half of the fine roots would have been initiated in warm soils with elevated N availability, and differences in fine-root N concentration still did not exist.
In a longer-term experiment conducted at Harvard Forest, MA (42° 28′ N, 72° 10' W), ecosystem fine-root respiration in warmed soil was equal to or less than that in the control treatment after 7 years of warming (Melillo et al. 2011). This response was not due to acclimation in specific fine-root respiration rates, but instead was due to a reduction in fine-root biomass (Zhou et al. 2011). We have not observed reductions in fine-root biomass in our sugar maple-dominated forests, but such a response remains a long-term possibility for further adjustment of ecosystem fine-root respiration, especially given enhanced N availability in the heated treatments. Our observed increase in growing season net N mineralization in the heat treatment relative to the control (68%, Table 2) is of similar magnitude to the increases observed in the Harvard Forest soil warming (Zhou et al. 2011, Butler et al. 2012), and reductions in root biomass are a common response to enhanced nutrient availability (Haynes and Gower 1995, Magill et al. 2004, Bakker et al. 2009, Jia et al. 2010) due to optimal partitioning of biomass for the acquisition of above- vs belowground growth-limiting resources (Bloom et al. 1985).
Our results for the first year of soil warming in sugar maple-dominated northern hardwoods indicate that mechanisms exist to reduce fine-root respiration to levels lower than those that would have occurred with an exponential increase in response to warmer temperature. Soil moisture availability had a large effect on treatment responses, with drier soil causing reduced respiration for all treatments. In the heat-only treatment, dry soil conditions and acclimation of specific respiration rates combined to reduce ecosystem fine-root respiration to rates that were nearly identical to the control. When soil moisture was enhanced by water additions in the water and heat + water treatments, respiration rates increased, but not enough to cause significantly greater CO2 efflux from ecosystem fine-root respiration. In the heat + water treatment, this was due to partial acclimation of specific respiration rates (Figure 5). Respiration rates for this treatment suggest that the trees would have to undergo further metabolic and/or physiological changes in order to remain C efficient in a future, warmer climate. Changes in fine-root-system biomass could play a role in such longer-term responses, but through the first full year of soil warming in sugar maple forests, such adjustments had not occurred.
Conclusion
A review of vegetation models commonly used as components of earth systems models indicates that only a few allow plant tissue respiration to acclimate to warmer conditions. Many allow respiration to increase exponentially with temperature and do not directly reduce root respiration in drier soil (Smith and Dukes 2013). As a result, model overestimates of the C flux associated with root respiration are likely, which could lead to underestimates of net primary productivity. In some of the models, respiration is constrained as a proportion of assimilated C, and this may provide a simple method to simulate the effects of respiratory acclimation, without the need to include complex mechanistic details. However, the fraction of gross primary production respired is known to vary widely among forest types and years, and in response to forest age and environmental conditions (DeLucia et al. 2007). Given the variation among our treatments in growing season fine-root respiration and the likelihood that the much drier soil in the heat-only treatment led to reduced photosynthesis, the proportion of assimilation that was allocated to root respiration in our study likely varied among treatments. Clearly, more detailed information on the potential range in proportions of assimilated C that can be allocated to root-system respiration and knowledge of the degree to which this can change in response to warmer and/or drier soil conditions is needed to ensure that we could adequately model the C flux involved using a proportion of assimilation approach. Including mechanistic algorithms for factors such as temperature acclimation and moisture effects in models would be a more ideal solution. Several recent efforts have attempted to do so for temperature acclimation, but it is clear that additional data from experimental studies are needed to adequately parameterize mechanistic responses to multiple global change factors (Smith and Dukes 2013).
Supplementary data
Supplementary data for this article are available at Tree Physiology Online.
Conflict of interest
None declared.
Funding
The research was funded by U.S. Department of Energy's Office of Science (BER) through the Midwestern Regional Center of the National Institute for Climatic Change Research (DE-FC02–06ER64158).
Acknowledgments
We would like to thank Jennifer Eikenberry, Kayla Griffith, Frederick Jarvi, Gerald Jondreau, Steve Rummel-Chadderdon, James Schultz and Dave Stimac for incredible assistance in the field and in the laboratory. We would also like to thank Rodney Chimner and John Hribljan for design assistance.
