Does manganese influence grass litter decomposition on a Hawaiian rainfall gradient? Authors: Elizabeth L Paulus, Peter M Vitousek Published: 2024 ABSTRACT AbstractPlant litter is a well-defined pool of organic matter (OM) in which the influence of manganese (Mn) on decomposition (both decomposition rate and the mix of compounds ultimately transferred to soil OM) has been clearly demonstrated in temperate forests. However, no similar study exists on grasslands and the effect of foliar Mn versus soil-derived Mn on litter decomposition is poorly known. We used a 5-month and 12-month field, and 10-month laboratory experiments to evaluate litter decomposition on the Kohala rainfall gradient (Island of Hawai‘i) in areas with different foliar and soil Mn abundances, and on which a single plant species (Pennisetum clandestinum) dominates primary production and the litter pool. The chemical imaging analyses of decomposed litter revealed that Mn2+ oxidized to Mn3+ and Mn4+ on grass litter during decompositions—hallmarks of Mn-driven litter oxidation. However, these transformations and Mn abundance did not predict greater litter mass loss through decomposition. These observations demonstrate that the importance of Mn to an ecosystem’s C cycle does not rely solely on the metal’s abundance and availability.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00442-024-05638-4.Keywords: Manganese, Organic carbon, Grassland, Redox cycling, Litter decomposition Introduction In natural and laboratory environments within temperate forests, manganese (Mn) abundance and availability correlate with litter decay (Cui and Dolphin 1990; Berg and McClaugherty 2003; Berg et al. 2007; Davey et al. 2007; Keiluweit et al. 2015; Whalen et al. 2018). A comprehensive literature review found Mn concentration so highly correlated with the degree of decomposition such that it appears to be “the single main factor” predicting OM decay in many sites (Berg et al. 2010). In this study, we investigate whether Mn influences grassland litter decomposition and whether the Mn source—foliar or soil—matters. Mn is a biologically essential element ubiquitous in soil. In natural environments, it exists in three oxidation states: Mn2+ is the most reduced, energetically stable, soluble species and the only nutritionally available (or bioavailable) form; Mn4+ is the most oxidized and least soluble form; and Mn3+, a powerful and unstable oxidant, is the most reactive form (van der Lee 1999; Rezanezhad et al. 2014; Keiluweit et al. 2016; Jones et al. 2018). Due to the thermodynamics of the single-electron oxidation or reduction step required to produce Mn3+ from Mn2+ or Mn4+, Mn3+ is an extremely unlikely intermediate between Mn2+ and Mn4+ (Luther 2005). If it is produced, Mn3+ can only persist if it is stabilized as a mineral or chelated by ligands; otherwise, it is rapidly oxidized to Mn4+ or reduced to Mn2+ (Madden and Hochella 2005; Webb 2005; Lan et al. 2017). Accordingly, the most efficient litter-degrading fungi, the white-rot fungi, produce a Mn-dependent enzyme (Mn peroxidases) to overcome the thermodynamic challenges Mn3+ poses and take advantage of its potency: Mn peroxidase oxidizes Mn2+ to Mn3+ to attack and degrade OM by depolymerizing complex molecular constituents of OM, especially lignin (Hofrichter 2002). Thermodynamics underlies why Mn peroxidase is the likely reason that decomposing litter accumulates oxidized Mn in the form of Mn3+/4+-oxides (Berg et al. 2010; Keiluweit et al. 2015). The plants take up Mn2+ from the rhizosphere through root cells, releasing H+ or low-molecular-weight organic acids to acidify the local environment and reduce Mn-oxides into bioavailable Mn (Rengel and Marschner 2005). In general, the acidic and/or reducing soil conditions favor Mn uptake in plants; basic and/or oxidizing soil conditions diminish Mn availability. Once absorbed, Mn is quickly transported through the xylem to photosynthetically active leaves via the transpiration stream (Page and Feller 2005). Consequently, Mn accumulates mostly in leaves but, due to its diverse biological roles, is found throughout the plant, mostly as Mn2+ (Alejandro et al. 2020). The plants do not resorb Mn during leaf senescence; as their leaves die and fall to the ground, they return Mn to surface soils, where decomposition oxidizes Mn2+ in litter to Mn3+/4+-oxides (Herndon et al. 2015; Keiluweit et al. 2015). Despite the tight recycling of plant-litter Mn, the plants can absorb Mn at 100–1000 times the rate at which they use the nutrient (Clarkson 1988); some species’ absorption rates mirror Mn2+ concentrations in soils, even to toxic levels (Marschner 1988; Fernando and Lynch 2015). Differences in species’ sensitivities to Mn toxicity is posited to have reshaped a temperate grassland that was treated with nitrogen (N) for a decade—N amendments acidified the soil, which increased Mn bioavailability to toxic levels and transformed the once forb- and grass-dominated community to exclusively grasses (forbs are more sensitive to Mn toxicity) (Tian et al. 2016); moreover, greater N additions correlated with greater Mn liberation and increased litter decomposition rates (Hou et al. 2021). Such shifts in litter biomass and decomposition rates suggest that Mn may shape an ecosystem’s C cycle. However, how Mn does so remains unclear, especially in grasslands, a largely overlooked ecosystem in Mn studies. The grasslands cover approximately one-quarter of Earth’s land surface, account for roughly one-third of net primary production on land, and store around 12% of total terrestrial C (Hoekstra et al. 2005; Janowiak et al. 2017). Their productivity makes them exceptionally useful to humans for food and forage production, and vulnerable to degradation and consequent C loss: 20–25% of grasslands have been degraded (Janowiak et al. 2017). We need to know how vulnerable or resilient grassland C stocks are to Mn-redox cycles. OM stability depends on the ecosystem, hinging on complex interactions among the abiotic and biotic factors of an environment (Schmidt et al. 2011; Lehmann and Kleber 2015). An ecosystem’s climate, potential organo-mineral interactions, nutrient accessibility, and plant and microbial community and functioning influence OM decomposition. Accordingly, Mn may not exercise the same control on litter decomposition in a grassland as it reportedly does in forests (Fujii et al. 2020; Santos and Herndon 2023). And the effect of foliar Mn versus soil-derived Mn on litter decomposition is poorly known. We use field and laboratory experiments to evaluate these dynamics on the Kohala rainfall gradient in areas with different foliar and soil Mn abundances, and on which a single plant species (kikuyu [Pennisetum clandestinum]) dominates primary production and the litter pool. Materials and methods Gradient sites The samples were decomposed in and collected from three sites on the Kohala Mountain rainfall gradient, a ~ 14-km transect on the Island of Hawai‘i that receives from < 300 to > 3200 mm mean annual precipitation (MAP) (Giambelluca et al. 2013); mean annual temperatures and elevation range from 23.5 °C and 50 m at the dry end to 16 °C and 1000 m at the wet end (Giambelluca et al. 2014). Although temperature, elevation, and rainfall vary along the Kohala gradient, previous studies on the gradient suggest that rainfall (and water availability) primarily underlie the emergent shifts in pedology and ecology (Chadwick et al. 2003; von Sperber et al. 2017; Vitousek et al. 2021). Three experimental sites that receive ~ 1578, 2163, or 3238 mm MAP were chosen because all three points are Andisols located on the same ~ 150,000-year-old Hawī volcanic formation (Chadwick et al. 2003; Sherrod et al. 2007), possess the same relief, are moderately acidic (Table 1), and have been grazed by cattle for over 100 years (Kagawa and Vitousek 2012). The field sites’ approximate locations are shown in Fig. S1. In all three sites, a single species of kikuyu grass (Pennisetum clandestinum) dominates primary production and (consequently) the litter pool; soil nutrients are relatively abundant here (Chadwick et al. 2007; Vitousek and Chadwick 2013), Basidiomycota DNA (the fungal group that includes ligninolytic white rot fungi) have been identified (Peay et al. 2017) and Mn peroxidase activity has been detected on this gradient (Paulus and Vitousek 2024). Each site also possesses distinct concentrations of Mn in its soil and in kikuyu grass (Table 1, Fig. S2): of these three sites, the driest site has the greatest soil and lowest grass Mn concentrations, the middle MAP site has moderate soil and the greatest grass Mn concentrations, and the wettest site has the lowest soil and moderate plant Mn concentrations. Table 1Mean annual precipitation (MAP), soil pH, soil manganese (Mn) concentrations, and grass Mn concentrations of the gradient sites used for the experimentSite NumberMAP (mm)Soil pHSoil pH SEMSoil Mn (µg Mn g−1 dried soil)Soil Mn SEMGrass Mn (µg Mn g−1 dried grass)Grass Mn SEM115785.3280.10312719250.673.084.989221634.4720.0181117917.80311.334.59332384.1500.0341644.13.570223.325.59MonthSite 1Site 2Site 3Rainfall (mm)Temperature (°C)Rainfall (mm)Temperature (°C)Rainfall (mm)Temperature (°C)Jan184.417.31248.716.14270.814.80Feb130.917.20166.915.98237.014.55Mar212.217.59286.616.35412.614.93Apr171.818.34237.116.97349.015.28May96.5019.09134.417.58227.315.82June109.020.07146.418.63267.316.58July116.120.44185.718.87324.416.91Aug97.9020.84135.419.37235.317.57Sept68.0020.6991.9019.44165.117.92Oct81.1020.48113.319.22183.617.63Nov145.819.27199.117.89282.216.27Dec163.917.82217.216.61283.414.92Mean131.519.10180.217.75269.816.10Median123.519.18176.317.74269.116.05Monthly temperature and rainfall data for field sites: site 1 (driest), site 2, and site 3 (wettest) (Giambelluca et al. 2013, 2014) Mean annual precipitation (MAP), soil pH, soil manganese (Mn) concentrations, and grass Mn concentrations of the gradient sites used for the experiment Monthly temperature and rainfall data for field sites: site 1 (driest), site 2, and site 3 (wettest) (Giambelluca et al. 2013, 2014) Soil and litter collection The soil and grass samples were collected in the summers of 2018 and/or 2019 and were measured in at least triplicate (n = 3–6). The soil samples were collected as continuous 10-cm cores from the A horizon, sealed in two layers of plastic bags, manually homogenized to remove root matter and rocks, and frozen at −20 °C. To determine percent soil moisture, subsamples were sieved to 2 mm, weighed, and oven-dried at 65 °C to constant weight. The grass collection was limited to aboveground litter. For chemical analyses, the grass litter subsamples were oven-dried at 65 °C and ground using a Wiley Mill. We determined background Mn concentrations in dried, ground soil and litter subsamples using X-ray fluorescence (XRF); background Mn data are presented in Fig. S2. In August 2021, freshly senesced kikuyu leaves were collected from the three of the gradient sites for the field and laboratory decomposition experiments. The leaf litter was air-dried, weighed, and separated into ~ 1 g aliquots. Each aliquot was sewn into a mesh litter bag (Fig. S3) and reweighed. Litter bags were constructed from 6 cm × 12 cm sections of 0.011-gauge fiberglass insect screen cloth folded in half lengthwise and secured by ribbon sewn onto three sides of the resulting 6-cm2 square. Litter-decomposition field experiment The field experiment was designed to parse the relative contributions of foliar versus soil Mn to litter decomposition under different rainfall regimes. Kikuyu dominates the plant communities, primary production and the litter pools in all three sites (1578, 2163, and 3238 mm MAP). We collected grass litter at each site, created grass litter bags, and distributed the litter bags among the three sites: three bags from each site were left to decompose at all sites for 5 or 12 months. Three additional bags per site were immediately preserved at 4 °C for initial chemical and synchrotron analyses. At the end of the 5- and 12-month decompositions, at least three bags of each type were collected from each decomposition site and preserved at 4 °C for 5- and 12-month-group analyses. We assessed mass loss, C content, Mn concentration, and distributions of Mn oxidation states in each litter bag to evaluate whether Mn concentration and availability correlate with litter decomposition and, if so, whether foliar or soil Mn better predicts the extent of litter decomposition. Manganese-treatment incubation experiment In addition to the field experiment, we used a laboratory decomposition experiment to directly test the effects of Mn abundance and availability on litter decomposition. We collected kikuyu litter from three gradient sites (receiving 1578, 2163, or 3238 mm MAP), air-dried the litter, and divided the litter into ~ 1 g aliquots per litter bag. We also subsampled litter for background/initial chemical analyses. To serve as substrate for the incubation, soil containing low background Mn (~ 644 µg Mn/g dried soil; Table 1) was collected from the wettest site. Three litter bags from each site were randomly assigned to one of two conditions: control or Mn treatment. Both conditions were maintained at 25 °C on a laboratory bench in equivalent un-lidded plastic tubs. For 10 months, the control and Mn-treatment conditions received weekly additions of 495 ml of deionized (DI) water (no additional Mn) or 17.6-µM manganese chloride (MnCl2), respectively, to approximate the average daily rainfall of the wettest site. These additions were dispersed uniformly across the soil substrate (and incubation container’s floor) (Fig. S3). Over the 10-month incubation, the Mn treatments provided ~ 2 mg total of supplemental Mn per g initial dry litter mass (0.35 g Mn m−2). At the end of the experiment, we harvested the litter to assess mass loss, changes in C and Mn contents, and distributions of Mn oxidation states across litter surfaces. Laboratory analyses Grass litter was air dried and weighed before and after decomposing to assess mass loss; initial samples of each experimental type were air dried and weighed immediately following collection. The subsets of each litter type were oven dried at 65 °C to constant weight to characterize air-dry-to-oven-dry mass conversions and for C content and total Mn analyses; C content was determined using a Carlo Erba NA 1500 Elemental Analyzer (EA), and Mn concentration was measured using inductively coupled plasma optical emission spectroscopy (ICP-OES), performed on a Thermo-Fisher ICAP 6300 Duo View. To measure soil pH, soil samples were mixed with DI water in 1 g:1 mL slurries and shaken for 5 min (Thomas 1996); slurries stood for 10 min before pH was read. Each soil pH is reported as an average of six measurements (Table 1). EA, ICP-OES. EA, ICP-OES, and XRF measurements were performed in the Environmental Measurements Facility (EMF) at Stanford University. Micro-X-ray absorption spectroscopy imaging Mn distribution and oxidation states on grass litter were determined using elemental maps and Mn micro-X-ray absorption near edge structure (µXANES) spectra. The samples were stored at − 20 °C between collection and synchrotron imaging. Micro-X-ray absorption spectroscopy (µXAS) imaging was performed on beamline 7–2 at the Stanford Synchrotron Radiation Lightsource (SSRL), SLAC National Accelerator Laboratory. This beamline uses a water-cooled, double-crystal Si(111) monochromator; the energy was calibrated using the first derivative of a Mn metal foil to 6537.7 eV (Manceau et al. 2012). Litter samples were sealed into sample holders using X-ray-transparent Kapton tape and polyethylene plastic wrap. Each litter sample was initially imaged using coarse resolution to map total Mn abundance. We used these data to choose regions for Mn multi-energy mapping and X-ray absorption near edge structure (XANES) spectroscopy to map Mn oxidation states across litter surfaces. Fine-resolution images were generated using a 25-µm beam focus on small areas of interest. We used principal component analyses (PCA) and simplex volume maximization (SiVM) in SMAK (v2.00) to choose the most appropriate locations for XANES spectroscopy (Webb 2011; Alfeld et al. 2017; Kravchenko et al. 2022). XANES spectra were normalized and verified in SixPack (v1.5.6) using previously published standards (Hansel et al. 2012; Johnson et al. 2016). To compare relative densities of Mn oxidation states on litter surfaces, we used SMAK to perform particle statistics and particle-density analyses. The particles were distinguished from backgrounds using InvBinary or Otsu thresholding algorithms, and the minimum-sized particle was defined as two pixels. The particle-density analyses identified particles in sample images as Mn2+, Mn3+, or Mn4+ following PCA and XANES analyses. The regions of Interest (ROI) were assigned by the Particle Statistics function in SMAK. Statistical and data analyses Means and standard errors of means were calculated for all soil analyses. We applied the Shapiro–Wilk test to assess normality in all datasets. When the data satisfied normality assumptions, we applied parametric tests: one way analysis of variance (ANOVA), two way ANOVA, and Tukey’s honest significant difference (HSD) post-hoc test, or the Student’s t-test. When the data did not meet normality assumptions, we applied nonparametric tests: the Kruskal–Wallis test and the Dunn-Bonferroni post-hoc test or the Wilcoxon–Mann–Whitney test. Significance was determined as p-values ≤ 0.05. Calculations, graphing, and statistics were performed in Microsoft Excel (v16.67) and R (v4.2.2). The synchrotron data were collected using SMAK (v2.00) (Webb 2011); spectra normalizations and statistics (principal component analyses [PCA] and linear Combination fitting [LCF]) were performed using Sixpack (v1.5.6) (Webb 2005). Kohala gradient mean annual precipitations (MAP) were reported from Giambelluca et al. 2013. Results Field litter-decomposition experiment To assess how litter Mn concentrations changed with decomposition, we compared Mn concentrations in initial and 5-month-decomposed litter collected from all three sites (Fig. 1A). The litter was too degraded and sparse to determine Mn concentrations in the samples that were decomposed in site 3 for 5 months and in all sites for 12 months. Initial litter Mn concentrations generally increased with rainfall: the Mn concentration in litter collected from the driest site (L1) was significantly lower than at the wettest site’s (L3); litter collected from the moderate rainfall site (L2) did not significantly differ from the other two sites. After 5 months, the greatest Mn concentrations were found in litter decomposed in site 1, with L1 and L3 showing significantly greater Mn concentrations than their counterparts decomposed in site 2. L1 decomposed in site 1 showed the greatest Mn concentrations of any group, initial or decomposed. Mn concentrations in litter decomposed in site 2 did not significantly differ across litter types (L1, L2, and L3). A Grass litter manganese (Mn) concentrations in initial and 5-month decomposed litter. Litter 1 (L1), litter 2 (L2), and litter 3 (L3) reflect litter collected from sites 1, 2, and 3. Hollow columns symbolize initial (freshly senesced) litter Mn concentrations; gray columns symbolize litter decomposition in site 1; and black columns symbolize decomposition in site 2. No Mn data are shown for 12-month decompositions or for site 3 5-month decompositions due to insufficient decomposed litter from these groups. B Particle density analyses of micro-X-ray absorption spectroscopy (µXAS) images (Figure S4). Columns indicate the relative percentages of Mn oxidation states found on the litter surfaces. White columns symbolize the percentage of Mn2+, gray columns Mn3+, and black columns Mn4+. Initial (freshly senesced) grass litter is designated by i. S1, S2, and S3 designate litter decomposed for 12 months in sites 1, 2, and 3. C Percent mass loss of litter over 5-month decomposition experiment; D percent mass loss of litter over 12-month decomposition experiment. L1, L2, and L3 show litter collected from sites 1, 2, and 3; S1, S2, and S3 designate litter decomposed in sites 1, 2, and 3. Letters above columns define statistically significant differences among litter types at p ≤ 0.05, n = 2–5. Error bars indicate standard error of the mean (SEM) µXRF imaging analyses reveal Mn compositions on the surface of litter shifted over 12-month decompositions (Fig. 1B); the µXRF images sourced for the particle-density analyses are shown in Fig. S4. The litter collected from Site 1 (L1) initially showed Mn2+ and Mn4+ on its surface. After 12 months, Mn2+ disappeared from the surfaces of L1 litter decomposed in sites 1 and 2, while Mn+3 appeared with Mn4+; litter decomposed in site 3 did not differ in Mn composition from the initial litter. The litter collected from site 2 (L2) also initially presented Mn2+ and Mn4+ on its surface. Mn4+ was the only oxidation state apparent on L2 litter decomposed in sites 1 and 2, while all three oxidation states appeared in the litter decomposed in site 3. Lastly, L3 litter initially presented almost exclusively Mn4+ and little Mn2+. Decomposed L3 litter showed Mn4+ dominated Mn compositions in litter decomposed in all three sites: Mn3+ contributed little to Mn compositions in litter decomposed in sites 1 and 2, and Mn2+ was absent in decomposed litter from all sites. We measured mass loss in all litter types to indicate degree of decomposition. After 5 months, litter types decomposed in site 3 generally showed the greatest mass loss (Fig. 1C); this relationship was statistically strongest for L2. The litter decomposed in Site 1 and 2 did not significantly differ in mass loss. 12-month decompositions show no consistent pattern (Fig. 1D). Linear analyses between 5-month-decomposition mass loss and litter Mn, soil Mn, and MAP reveal the only positive association between mass loss and MAP (Fig. 2); litter Mn and soil Mn showed negative linear relationships. The difference between R2 values for 5-month- and 12-month-decompostion % mass loss versus MAP (0.63 and 0.09, Fig. 2E, F) confirms that MAP reliably predicts % mass loss for the 5-month decomposition but not for the 12-month decomposition. Accordingly, MAP and not Mn (whether foliar or soil) correlates with litter decomposition in the field. A Variation in percent mass loss of litter at 5 months versus litter Mn concentration, B at 12 months versus litter Mn concentration, C at 5 months versus soil Mn concentration, D at 12 months versus soil Mn concentration, E at 5 months versus mean annual precipitation (MAP), and F at 12 months versus MAP. Dotted lines show linear trendlines for each plot; the equation and coefficient of determination (R2) for each line is shown in the top right corner of each plot We evaluated carbon-to-nitrogen ratios (C/N) as another metric of litter decomposition over 5 months and 12 months (Fig. S5); decreased C/N in litter generally indicates greater decomposition. We found no statistically significant differences among the permutations of litter type and decomposition site for either the 5-month or 12-month field experiment. Manganese-treatment incubation experiment Particle density analyses of litter from the incubation experiment reveal distinct patterns in Mn oxidation states on litter surfaces (Fig. 3); source µXAS images of litter are shown in Fig. S6. Mn treatment significantly increased Mn2+ in the Mn-treated litter from Site 2 but no other site (Fig. 3A, D); Mn3+ and Mn4+ prevalence did not change with Mn treatment (Fig. 3B–D). To isolate the effect of Mn availability on litter Mn oxidation states, we compared the aggregated control versus Mn treatment data (Fig. 3E, F). Paired analyses revealed that Mn treatment significantly increased the percentages of Mn2+ and Mn3+ but not Mn4+ on decomposed litter (Fig. 3E). These data confirm that Mn treatment increased Mn bioavailability at the surface but not necessarily Mn redox cycling. Micro-X-ray absorption spectroscopy (µXAS) analyses of litter types (A–D) and aggregate litter per condition (E, F); source µXAS images are shown in Figure S6. Element concentrations are measured as photon counts per image pixel, or the signal intensity of the target oxidation state of manganese (Mn) in the X-ray images’ regions of interest. A White boxes and circles represent Mn2+ percentages, B gray boxes and circles Mn3+, and C black boxes and circles Mn4+. C1, C2, and C3 are the control litter groups, and MnT1, MnT2, and MnT3 are the Mn-treatment litter groups; numbers (1, 2, and 3) indicate the respective litter-collection site. D, F White columns represent Mn2+ percentages, gray columns Mn3+, and black Mn4+. E Hollow boxes and circles represent litter in the control condition, and gray boxes and circles show litter under Mn treatment. In A–C and E, the middle line in each box represents the median, the bottom and top lines represent 1st and 3rd quartiles, the x represents the mean, and the whiskers extend to the minimum and maximum values; outliers are shown as circles. In D and F error bars indicate standard error of the mean (SEM). Asterisks define significant differences A–D among litter types or E, F between treatments at p ≤ 0.05; brackets show paired comparisons Whether analyzed according to litter type or aggregate condition, no significant differences were found in mass loss between the control and Mn treatment groups (Fig. 4A, B). Similarly, no differences in C/N ratios were found between the control and Mn treatment groups, whether divided by litter type or examined in aggregate (Fig. S7A, B). Accordingly, Mn treatment (increased Mn abundance and availability) did not increase litter decomposition over the 10-month laboratory incubation experiment. Percent mass loss in litter incubated for 10 months. Litter 1 (L1), litter 2 (L2), and litter 3 (L3) reflect litter collected from sites 1, 2, and 3. Hollow columns and black-filled columns represent control and manganese (Mn) treatment litter, collected from sites 1 (the driest), 2, and 3 (the wettest). Hollow boxes and black-filled boxes represent aggregated control and Mn-treatment litter, respectively. Control and Mn-treatment litter are compared by A litter type (collection site) and by B aggregated condition. No significant differences were found A among litter type or B between condition (p ≤ 0.05, n = 3 per litter type, n = 9 per condition). Error bars indicate standard error of the mean (SEM). Boxplots show means (X) and inclusive quartiles Discussion We hypothesized that Mn bioavailability would correlate with litter decomposition in the field and laboratory experiments. While a 6-year decomposition time was used for pine needles in a temperate forest (Keiluweit et al. 2015), we used shorter study periods (5, 10, and 12 months) because we expected to see rapid litter decomposition in the Kohala grassland due to the sites’ nutrient availability (Chadwick et al. 2007; Schmidt et al. 2011; Vitousek and Chadwick 2013; Lehmann and Kleber 2015), the activity of the ligninolytic enzyme Mn peroxidase in the sites (Paulus and Vitousek 2024), and grass’s low lignin content (relative to pine needles’) (Meentemeyer 1978; Melillo et al. 1982; Wedin and Tilman 1990; Talbot et al. 2012; Santos and Herndon 2023). We observed differences in Mn content and shifts in Mn oxidation states on litter as it decomposed in the field and laboratory, suggesting Mn could mechanistically degrade litter; but we found no evidence that Mn governed litter decomposition. In the field experiment, Mn accumulated in all litter that decomposed for 5 months in Site 1 relative to freshly senesced litter collected along the gradient (Fig. 1A); Mn content did not change in litter decomposed in site 2 and there was not enough material to measure Mn in grass decomposed in site 3. µXAS analyses of litter surfaces demonstrated that the composition of Mn oxidation states evolved from Mn2+ and Mn4+ on freshly senesced grass leaves to Mn2+, Mn3+, and Mn4+ on 12-month-decomposed litter (Fig. 1B; Fig. S4). Mn2+ and Mn3+ mark the early stages of decomposition; as decomposition progresses, Mn oxidizes into Mn3+/4+, especially in hotspots of OM breakdown (Thompson et al. 2005; Hansel et al. 2012; Herndon et al. 2014; Keiluweit et al. 2015). We expected the greater Mn concentration and oxidation to correlate with greater litter decomposition, evidenced by increased mass loss and decreased C/N ratios. However, mass loss correlated with MAP (Fig. 2), not with Mn content, and there were no significant differences among C/N ratios among the litter types decomposed for 5 or 12 months (Fig. S5). Despite controlling for water availability and temperature, similar patterns emerge from the laboratory incubation study. Over the 10-month incubation, the Mn treatments provided a total of ~ 2 mg of available supplemental Mn per g initial dry litter mass. This concentration of supplementary Mn was three times the substrate soil’s Mn concentration and 7–29 times the grass litters’ Mn concentrations (Table 1). This additional Mn was sufficient to observe the effect of supplementary Mn availability without risking Mn toxicity to the fungal community decomposing the litter: Mn treatment successfully supplemented the availability of Mn2+, which translated to greater total Mn3+ in the Mn treatment group in comparison to the control (Fig. 3E; Fig. S6); fungal hyphae were also visible in Mn treated litter bags and not in control litter bags (Fig. S3). The primary mechanism through which Mn is posited to degrade litter is through the ligninolytic enzyme Mn peroxidase. Given that lignin generally composes 5% of grassland litter versus 19% of temperate forest litter, our Mn application was comparable on a per g lignin basis to previous studies that investigated the effect of Mn on temperate forests’ litter decay (Meentemeyer 1978; Melillo et al. 1982; Wedin and Tilman 1990; Gholz et al. 2000; Whalen et al. 2018; Zhang et al. 2024). But these differences in Mn content, Mn oxidation states, and apparent fungal activity did not correspond with greater mass loss (Fig. 4) or decreased C/N ratios (Fig. S7) in the Mn-treated litter. We did not see Mn treatment enhance litter decomposition over the 10-month decomposition experiment. Although, in the current study, we saw evidence of Mn oxidation and fungal activity (Fig. S6 and Fig. S3C–F)—the primary mechanism through which Mn may destabilize organic matter—and previous studies discovered evidence of Basidiomycota and Mn peroxidase activity in the experimental sites (Peay et al. 2017; Paulus and Vitousek 2024), we found no correlation between Mn content (foliar, soil, or total) and grass litter decomposition. Even when water availability and temperature (proxies for MAP in the field study) were controlled in the incubation experiment, Mn abundance and availability did not determine litter decay. Previous studies that observed links between Mn and litter loss focused on spruce and needle litter from temperate forests, following their decay over longer experimental periods than in the present study (Berg 2000; Davey et al. 2007; Keiluweit et al. 2015; Berg et al. 2015). Whaley et al. (2018) also studied needle litter under Mn amendment but limited their experiment to 6 months; as in our study, they found Mn addition did not increase total litter mass loss. The authors note that their study could have been too short, proposing that Mn amendments would have increased litter mass loss during late-stage decay rather than early-stage because the relationship between Mn and litter decay strengthens as litter decay proceeds (Berg et al. 2007). It is unlikely, however, that our experiments were too short to observe Mn influence decay. In the field study, most litter bags lost ≥ 50% mass. So little litter remained in the litter bags after 12 months that it was difficult to distinguish grass litter from incidental matter that might have washed through the screen with rainfall and remained caught in the bag; this was especially true of the litter bags that decomposed in site 3 (the wettest site). The laboratory study litter bags saw comparable mass losses of ~ 50%. Furthermore, the shifts in Mn oxidation states, from reduced to oxidized, suggests the litter samples experienced advanced stages of breakdown. The litter decayed, but not as predicted by Mn availability and oxidation state. Accordingly, Mn does not control grass litter decomposition in this ecosystem. While Mn availability may be the “single main factor” predicting litter decomposition in temperate forests (Berg et al. 2010), it does not appear to wield the same influence in grasslands—despite evidence of Mn peroxidase and fungal activity in the incubation and field experiments. Why Mn has distinct influences in grasslands and forests is not clear. In forest soils, Mn peroxidase appears to perform most efficiently around pH 4.5–5 (Fujii et al. 2013, 2020). The Kohala sites’ soil pH range from 4.1520–5.328 (Table 1); this evidence and the Mn peroxidase activity measured in these sites suggests that the Mn peroxidase produced here should be viable. But perhaps the Basidiomycetes in these sites do not produce enough of this enzyme (within the suite of decomposition enzymes they produce) for Mn to determine decomposition rate. Further study on Mn-dependent enzyme activity and expression is warranted in grasslands to investigate if these findings apply broadly to grasslands. In forests, foliar Mn concentration increases with lignin content and decomposition rate, and high Mn/lignin ratios correlate with greater Mn peroxidase activities (Fujii et al. 2020; Santos and Herndon 2023). Grass litter contains less lignin than forest litter (e.g., pine needles) does, perhaps indicating one reason why Mn does not predict litter decay in the grassland as it seems to in forests. The distinct substrate chemistries in grasslands versus forests do not alone explain the difference; they reflect their ecosystems’ distinct environments and microbial-community functioning that regulates decomposition (Schmidt et al. 2011; Lehmann and Kleber 2015). Future studies should more deeply interrogate these microbial communities, their activities, and the environmental factors (e.g., redox conditions) that shape both. Mn availability and oxidation state did not predict grass litter decomposition—either along a grassland rainfall gradient under field conditions or in Mn-treated incubations. These findings suggest that Mn is cycling in the grassland at capacity—the ecosystem’s microbial-community composition and/or resource availability cannot support faster or greater Mn cycling, rendering supplemental Mn inconsequential to litter decomposition. The importance of Mn to an ecosystem’s C cycle may only emerge as other environmental factors converge to otherwise retard the rate of decay. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 14089 KB) Acknowledgements We thank Parker Ranch, Ponoholo Ranch, and Kamehameha Schools for access to the research sites; Ulu Mau Puanui for equipment support; Kehaulani Marshall and Healohamele Genovia for field assistance; Doug Turner and Dr. Guangchao Li for laboratory analyses; Drs. Juan Lezama, Kristin Boye, and Sam Webb for SLAC SSRL synchrotron data collection and interpretation; and Drs. Scott Fendorf, Oliver Chadwick, Karen Casciotti, and Kabir Peay of Stanford University for their invaluable input on this work. Author contribution statement ELP and PMV designed the study and conducted the fieldwork. ELP conducted the laboratory analyses, analyzed the data, and wrote the article with contributions from PMV. Funding This work was supported by United States National Science Foundation Award Number 2027290, in collaboration with Dr. Mengqiang Zhu (University of Wyoming). ELP was also supported by the Department of Energy Science Graduate Student Research Fellowship (DOE SCGSR). Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Code availability The code used during the current study are available from the corresponding author on reasonable request. Declarations Conflict of interest The authors declare that they have no conflict of interest nor competing interests. Ethical approval NA. Consent to participate NA. Consent for publication NA. References Associated Data Supplementary Materials Supplementary file1 (DOCX 14089 KB) Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. The code used during the current study are available from the corresponding author on reasonable request.