Symbiotic nitrogen (N) fixation has the potential to replenish fire-induced N losses in frequently burned ecosystems. A strong relationship between fire and fixation may exist because fire volatilizes N and mineralizes phosphorus (P), creating N-poor, P-rich soils that favor plants capable of N-fixation. However, human activities have enriched ecosystems with N, which may complicate the interplay among fire, fixation, and soil P. We evaluated how N and P modulate the relationship between fire and symbiotic N fixation in longleaf pine savannas, where it was previously documented that N fixation fails to replenish N losses from fire. Across gradients of stand age and fire frequency, we investigated how N and P availability influence fixation, and we established a nutrient addition experiment to evaluate the effects of N and P on legume growth, fixation, and mycorrhizal investment. We uncovered a clear signal of P limitation of herbaceous legumes. Legume growth and fixation were linked to the availability of soil mineral P and were further stimulated by P additions. In contrast, neither soil N availability nor N additions affected legume growth or fixation.

Synthesis:

Our findings suggest that symbiotic N fixation in sandhill longleaf pine savannas is controlled by soil P availability, which varies according to soil age and parent material. Therefore, fixation may only counterbalance N losses from fire if there is enough P in the soil to support this process. Nonetheless, recent N enrichment in contemporary longleaf pine ecosystems may have reduced the importance of N fixation as a post-fire recovery mechanism.

Study site         

We conducted this research at Fort Moore Military Installation in southwestern Georgia (previously known as Fort Benning, and hereafter referred to as Moore) and Eglin Air Force Base in northwestern Florida (hereafter, Eglin), which respectively represent the Fall-Line and Coastal Plain ranges of sandhill longleaf pine savanna ecosystems (Peet, 2006). We received permission to conduct fieldwork through contract RC-2328 with the Strategic Environmental Research and Development Program. The climate at both sites is humid subtropical, and precipitation occurs throughout the year with a wetter season from May to September (MAT = 18.0˚ and 18.7˚C and MAP = 1,260 and 1,802 mm at Moore and Eglin, respectively).

Both military installations were acquired by the Department of Defense in the 1930s and have a mixed history of land use including agriculture, silviculture, grazing, and fire exclusion (Frost, 1993; Rodgers & Provencher, 1999), but now support large tracts of land that are managed with regular, low-intensity prescribed fire to promote longleaf pine restoration and conservation (mean ± SE fire return intervals of our research plots are 2.5 ± 0.1 and 6.2 ± 1.1 years, at Moore and Eglin respectively). These prescribed fires are intended to mimic the historical fire regime (Van Lear et al., 2005) and maintain an open-canopy longleaf pine overstory with pyrogenic grasses and forbs in the understory. Scrub oaks (especially Quercus laevis Walt.) are common in the midstory. All study plots were established on Lakeland sands (thermic, coated Typic Quartzipsamments), and have similar soil properties, including soil texture, structure, and carbon (C) and N content, which we determined in a prior study (Tierney et al., 2019; Table S1).

In 2014, we established 54 1-ha plots of longleaf pine distributed across two gradients at Moore and Eglin: fire return interval (1.5 – 20 years, calculated from 20 years of prescribed and natural fire events) and stand age (2-227 years old). These included 18 planted and 10 naturally regenerated (“mature”) stands at Moore and 18 planted and 10 naturally regenerated stands at Eglin. Plots were randomly positioned within plantations and burn units. Many of the plantations were previously mixed-pine forests that were harvested and site prepped (drum chopping sometimes paired with herbicide application) prior to planting with longleaf seedlings. Stand ages of naturally regenerated stands were estimated with tree ring analysis of the oldest tree (Tierney et al., 2019).

At random locations within each 1-ha plot, we established either two, three, or four 100-m² subplots in which we conducted soil sampling and legume censuses. These subplots were established to capture within-hectare variability, but due to logistical constraints, we could only establish the maximum number of subplots (4) in 16 of the 54 larger study plots.

Observational study

Soil P and N availability

We sampled mineral soils to determine phosphate availability with both the Mehlich-III method and the anion-resin method, which measure two different pools of mineral phosphate with different levels of biological availabilities. While anion resins are well-suited to measure the more labile pool of phosphate that closely mimics P availability in the rhizosphere (Robertson et al. 1999), Mehlich-III additionally extracts a potentially less-available pool of phosphate that is adsorbed to soil colloids and may not be immediately available for plant uptake (Mehlich 1984). In each subplot, we collected 6 soil cores (2.5 cm diameter and 20 cm depth) and homogenized them by depth increments (0-10 cm and 10-20 cm). We recorded the time passed between a fire event and soil sample collection, which ranged between 51 days and 10.8 years. Samples were removed of rocks, roots and organic matter and transported in coolers back to our laboratory at the University of Georgia for immediate analysis. Further details of the laboratory methods to determine Mehlich-III and resin-extractable P are provided in the Supplemental Methods.

To quantify N availability, we assessed N mineralization per subplot at a 20 cm depth with four sequential extractions (December 2015 to April 2016) using buried bag technique with KCl extractions (Robertson et al., 1999; Tierney et al., 2019). Details of the field and laboratory methods for this analysis are also provided in the Supplemental Methods.

Soil extracellular enzymes

We measured the activity of two extracellular enzymes important to the liberation of P and N from organic matter in soils – acid phosphatase (AP) and leucyl aminopeptidase (LAP). AP activity was determined using methylumbelliferyl-linked substrates and LAP activity with methylcouramin-linked substrates as previously described (Bell et al., 2013), and enzyme activities were expressed per g of dry soil. Soils were collected as described above and were kept at 4˚C until analysis, which we performed within two weeks of sample collection. All soil sampling for the determination of phosphate availability, N mineralization, and enzyme activity was completed before applying nutrient-addition treatments.                          

Nutrient-addition experiment

From our 54 plots of longleaf pine, we selected 20 plots to establish a nutrient-addition experiment (9 plots at Moore and 11 at Eglin). Each selected plot contained at least three 100-m² subplots in which we had conducted the observational study that would also serve as “treatment” subplots (hereafter treatment plots). Of these 20 plots, four were plantations between 3 and 4 years of stand age, four between 11 and 15 years, four between 19 and 23 years, and 8 stands were naturally regenerated (67 to at least 227 years old; Fig S1a). These 20 plots also varied in their fire return interval and time since fire (Fig. S1). In each plot, each of the three treatment plots were randomly assigned one of three treatments (hereafter treatment plots): N addition (50 kg N ha^-1 year^-1, ammonium nitrate fertilizer), P addition (25 kg P ha^-1 year^-1, triple super phosphate fertilizer) or control (no fertilizer addition).

Legume biomass, N fixation, mycorrhizal colonization, and tissue chemistry

We estimated legume biomass in each treatment plot by pairing census data with species-specific allometric equations (Table S2). Two pre-treatment censuses (2015-2016) and one post-treatment census (2017) were conducted in each subplot at the end of the growing season (July-August) in which we identified legumes to species and counted the number of individuals, stems, and leaves of each species present. Then, we paired abundance and size measurements with site- and species-specific allometric equations to determine legume aboveground biomass in each subplot (Tierney et al. 2019). We calculated the change in biomass as the difference between the average pre-treatment legume biomass (2015-2016) and post-treatment biomass (2017) in each subplot.

In July 2017, after 4 months of fertilization, we excavated the root systems of herbaceous legumes in each subplot. Overall, we excavated 524 individuals total from 23 species of herbaceous legumes to determine the prevalence of root nodules, which are the structures that house symbiotic N-fixing bacteria. Specifically, we targeted 14 to 35 individuals from nine of the most abundant species at each plot, for which we had previously determined species-specific N fixation rates. These species represented 87.9 and 88.3% of the total legume biomass at Moore and Eglin, respectively (Tierney et al. 2019). From 106 individuals growing within treatment plots, we collected live nodules and determined nodule mass-specific N fixation rates using acetylene reduction assays according to the methods of Tierney et al. (2019), and these data were included in allometric equations estimating biomass-based nitrogen fixation rates for these species (Tables S2 and S3). A conversion factor of 4.8 µmol ethylene: µmol N₂, which was determined by a greenhouse experiment of the same species (Ament et al., 2018), was used to convert rates of acetylene reduction to rates of N fixation.

To determine mycorrhizal colonization of legume roots, we refrigerated the fine roots of the legumes in 60% ethanol until they could be cleared with 10% KOH (65°C for 6 h) and stained with 0.05% trypan blue (70°C for 15 min) following the methods of Wurzburger and Wright (2015). We quantified mycorrhizal colonization as the percentage of root length containing mycorrhizal structures using the random intercept method detailed by McGonigle et al. (1990). We only counted arbuscules in this measure, as they are the mycorrhizal structures purportedly involved in the trading of C and nutrients (Hawkins et al., 2023).

We collected legume foliage in each treatment plot, including 68 individuals from 12 species, to determine how nutrient addition treatments may alter foliar nutrition and stoichiometry, which can indicate nutrient limitation in legumes (Batterman et al., 2013; Ren et al., 2017). We collected leaves only from the species that were present in all three treatment subplots within each of the 20 study plots to minimize the potentially confounding effects of plot-by-species interactions (Table S4). To quantify foliar chemistry, we dried the foliage at 70° C for 48 h, ground it to a fine powder, and determined tissue N content by combustion analysis and P content by acid digestion (D’Angelo et al., 2001; Greweling, 1976).

References:

Ament, M. R., Tierney, J. A., Hedin, L. O., Hobbie, E. A., & Wurzburger, N. (2018). Phosphorus and species regulate N2 fixation by herbaceous legumes in longleaf pine savannas. Oecologia, 187(1), 291–290. https://doi.org/10.1007/s00442-018-4129-z

Batterman, S. A., Wurzburger, N., & Hedin, L. O. (2013). Nitrogen and phosphorus interact to control tropical symbiotic N2 fixation: A test in Inga punctata. Journal of Ecology, 101(6), 1400–1408. https://doi.org/10.1111/1365-2745.12138

Bell, C. W., Fricks, B. E., Rocca, J. D., Steinweg, J. M., McMahon, S. K., & Wallenstein, M. D. (2013). High-throughput fluorometric measurement of potential soil extracellular enzyme activities. Journal of Visualized Experiments : JoVE, 81, e50961. https://doi.org/10.3791/50961

D’Angelo, E., Crutchfield, J., & Vandiviere, M. (2001). Rapid, sensitive, microscale determination of phosphate in water and soil. Journal of Environmental Quality, 30(6), 2206–2209. https://doi.org/10.2134/jeq2001.2206

Frost, C. C. (1993). Four centuries of changing landscape patterns in the longleaf pine ecosystem. Proceedings of the Tall Timbers Fire Ecology Conference, 18, 17–43.

Greweling, T. (1976). Chemical analysis of plant tissue. Agronomie, 6, 1–35.

Hawkins, H.-J., Cargill, R. I. M., Van Nuland, M. E., Hagen, S. C., Field, K. J., Sheldrake, M., Soudzilovskaia, N. A., & Kiers, E. T. (2023). Mycorrhizal mycelium as a global carbon pool. Current Biology, 33(11), R560–R573. https://doi.org/10.1016/j.cub.2023.02.027

McGonigle, T. P., Miller, M. H., Evans, D. G., Fairchild, G. L., & Swan, J. A. (1990). A new method which gives an objective measure of colonization of roots by vesicular—Arbuscular mycorrhizal fungi. New Phytologist, 115(3), 495–501. https://doi.org/10.1111/j.1469-8137.1990.tb00476.x

Peet, R. K. (2006). Ecological classification of longleaf pine woodlands. In The Longleaf Pine Ecosystem (pp. 51–93).

Ren, F., Song, W., Chen, L., Mi, Z., Zhang, Z., Zhu, W., Zhou, H., Cao, G., & He, J.-S. (2017). Phosphorus does not alleviate the negative effect of nitrogen enrichment on legume performance in an alpine grassland. Journal of Plant Ecology, 10(5), 822–830. https://doi.org/10.1093/jpe/rtw089

Robertson, G. P., Coleman, D. C., & Bledsoe, C. S. (1999). Standard Soil Methods for Long-Term Ecological Research. Oxford University Press. http://ebookcentral.proquest.com/lib/princeton/detail.action?docID=241674

Rodgers, H. L., & Provencher, L. (1999). Analysis of longleaf pine sandhill vegetation in Northwest Florida. Castanea, 64(2), 138–162.

Tierney, J. A., Hedin, L. O., & Wurzburger, N. (2019). Nitrogen fixation does not balance fire-induced nitrogen losses in longleaf pine savannas. Ecology, 100(7), e02735. https://doi.org/10.1002/ecy.2735

Van Lear, D. H., Carroll, W. D., Kapeluck, P. R., & Johnson, R. (2005). History and restoration of the longleaf pine-grassland ecosystem: Implications for species at risk. Forest Ecology and Management, 211(1), 150–165. https://doi.org/10.1016/j.foreco.2005.02.014

Wurzburger, N., & Wright, S. J. (2015). Fine‐root responses to fertilization reveal multiple nutrient limitation in a lowland tropical forest. Ecology, 96(8), 2137–2146. https://doi.org/10.1890/14-1362.1