Thermal responses of dissolved organic matter under global change

The diversity of intrinsic traits of different organic matter molecules makes it challenging to predict how they, and therefore the global carbon cycle, will respond to climate change. Here we develop an indicator of compositional-level environmental response for dissolved organic matter to quantify the aggregated response of individual molecules that positively and negatively associate with warming. We apply the indicator to assess the thermal response of sediment dissolved organic matter in 480 aquatic microcosms along nutrient gradients on three Eurasian mountainsides. Organic molecules consistently respond to temperature change within and across contrasting climate zones. At a compositional level, dissolved organic matter in warmer sites has a stronger thermal response and shows functional reorganization towards molecules with lower thermodynamic favorability for microbial decomposition. The thermal response is more sensitive to warming at higher nutrients, with increased sensitivity of up to 22% for each additional 1 mg L-1 of nitrogen loading. The utility of the thermal response indicator is further confirmed by laboratory experiments and reveals its positive links to greenhouse gas emissions.


Supplementary Figures
Figure S1.Study area on the three mountainsides across the contrasting climatic zones for field microcosm experiments (a).For the subtropical wet climate, we selected the Laojun Mountain in the southeastern margin of the Tibetan Plateau, China (red triangle).For the temperate arid climate, we selected the Dangjin Mountain in the northern Tibetan Plateau, China (green triangle).For the subarctic climate, we selected the Balggesvarri Mountain in Northern Europe, Norway (yellow triangle).On each mountainside, we selected five to six different elevations.The elevations were 3,822, 3,505, 2,915, 2,580 and 2,286 m a.s.l. on Laojun Mountain (red dots; b), 3,905, 3,477, 3,050, 2,670, 2,214 and 1,750 m a.s.l. on Dangjin Mountain (green dots; c), and 750, 550, 350, 170 and 20 m a.s.l. on Balggesvarri Mountain (yellow dots; d).Sources: GMTED2010 layer at 7.5 arc-sencond resolution (https://www.usgs.gov/)and GloRic version 1.0 layer at 15 arc-second resolution (https://www.hydrosheds.org/page/gloric).Features included the percentage of statistically significant molecular thermal responses (MERs, %), and the lowest, mean and median MERs.MERs were quantified with the Spearman's correlation coefficient ρ between the relative abundance of each DOM molecule and water temperature.We categorized all molecules into 100 bins (that is, 1% of molecules per bin) according to the magnitude of their positive or negative MERs.We sequentially removed the molecules from the bins with lower magnitude of MERs and then calculated molecular features of the remaining molecules.We categorized all molecules into 100 equal-sized bins (that is, 1% of molecules per bin) according to the magnitude of their positive or negative MERs.We sequentially removed the molecules from the bins with lower magnitude of MERs and generated 100 removal scenarios.The measurements of molecular traits were visualized by plotting against the lowest MERs of remaining molecules for each removal scenario.Blue and red dashed lines highlight the cutoff of 100% significantly negative and positive MERs for the remaining molecules, respectively.At a given removal scenario, we also considered other molecular features such as the percentage of significant responses (%), the percentage of molecules removed, and mean and median MERs (see Fig. S3).Data are presented as the means ± s.e averaged across molecules for each removal scenario.In the first removal scenarios, n = 1,431, 3,829 and 1,417 molecules for the warm-depleting molecules in the three climate zones, respectively; n = 979, 1,620 and 914 molecules for the warm-accumulating molecules in the three climate zones, respectively.These n numbers were then sequentially reduced by removal of one bin of molecules and not provided here considering hundreds of numbers.added into microcosms and mixed with artificial lake water.We simulated the dynamics in beakers in laboratory with the same sediments and artificial lake water.It should be similar to the real field experimental conditions regarding the sediment settlement.We mixed the sediments with water is because we expected the sediments to be evenly distributed in the bottom of microcosms at the initial state of experiments.We found it would take at least one day for the sediments to settle down in the system.The relationships between the iCERs calculated using all molecules and significant MER molecules.The significance was examined using one-sided F-statistics.In the first removal scenarios, n = 1,431, 3,829 and 1,417 molecules for the warm-depleting molecules in the three climate zones, respectively; n = 979, 1,620 and 914 molecules for the warm-accumulating molecules in the three zones, respectively.These n numbers were then sequentially reduced by removal of one bin of molecules and not provided here considering hundreds of numbers.The area colors from grey to red in (b) represent compound classes with different GFE varying from high to low.

Figure S2 .
Figure S2.The distribution of DOM molecules with positive (red) or negative (blue) molecular thermal responses (MERs) in van Krevelen diagrams in the subtropical wet (Laojun Mountain, China), temperate arid (Dangjin Mountain, China) and subarctic (Balggesvarri Mountain, Norway) climate zones.MERs were quantified with the Spearman's correlation coefficient ρ between the relative abundance of each DOM molecule and water temperature.Each molecule is colored by its MER.

Figure S3 .
Figure S3.Molecular features along the gradient of the percentage of molecules removed.

Figure S4 .
Figure S4.The gradual changes in molecular traits along the continuum of negative (blue) to positive (red) molecular thermal responses (MERs) in the subtropical wet (Laojun Mountain, China), temperate arid (Dangjin Mountain, China) and subarctic (Balggesvarri Mountain, Norway) climate zones.We considered molecular traits of modified aromaticity index (AIMod), H/C ratio, O/C ratio, and nominal oxidation state of carbon (NOSC).We

Figure S5 .
Figure S5.Violin plots of Gibbs free energy (GFE, a) and H/C ratio (b) across compound classes of DOM.Smaller dots are the values for individual molecules; minima and maxima are the top and bottom of the violin, and the black triangles show the mean values for all molecules within each compound class.Aliphatics, n = 4,702; peptides, n = 4,161; low O unsaturated, n = 9,445; fatty acids & carbohydrates, n = 1,306; aromatics, n = 4,351; condensed aromatics, n = 2,422; high O unsaturated, n = 3,113 molecules.The violin colors from grey to red represent compound classes with different GFE varying from high to low values.

Figure S6 .
Figure S6.Violin plots of the indicator of compositional-level thermal responses (iCER) calculated with statistically significant (P ≤ 0.05) molecular thermal responses (MERs) in the subtropical wet (Laojun Mountain, China), temperate arid (Dangjin Mountain, China) and subarctic (Balggesvarri Mountain, Norway) climate zones.Smaller dots are the iCER for individual samples; minima and maxima are the top and bottom of the violin, and the black triangles show the mean values for each region (i.e., climate zone).Subtropical wet, n = 150; temperate arid, n = 180; subarctic, n = 150 samples.

Figure S7 .
Figure S7.The elevational patterns in the indicator of compositional-level thermal responses (iCER) of DOM across 10 nutrient levels in the subtropical wet (Laojun Mountain, China), temperate arid (Dangjin Mountain, China) and subarctic (Balggesvarri Mountain, Norway) climate zones.The iCER values were calculated with statistically significant (P ≤ 0.05) molecular thermal responses (MERs).We plotted iCER against elevation for each nutrient level and statistical significance of linear model fits with onesided F-statistics is indicated by solid (P ≤ 0.05) or dotted (P > 0.05) lines.

Figure S8 .
Figure S8.Total organic carbon (TOC) concentrations of sediments along the elevational (left panel) and nutrient (right panel) gradients (a), and their relationships with the indicator of compositional-level thermal responses (iCER, b) in the subtropical wet (Laojun Mountain, China), temperate arid (Dangjin Mountain, China) and subarctic (Balggesvarri Mountain, Norway) climate zones.The relationships in (a) were visualized with loess regression models.Blue dotted lines indicate the initial sediment TOC in the subtropical wet and subarctic climate zones, and orange dotted lines indicate the initial sediment TOC in the temperate arid climate zone.We plotted the iCER against sediment total organic carbon for all nutrients (black line) and each nutrient level (colored lines).The relationships in (b) are indicated by solid (P ≤ 0.05) and dotted (P > 0.05) lines estimated using linear models with one-sided F-statistics.

Figure S9 .
Figure S9.The principal component analysis (PCA) shows relationships among iCER and weighted means of formula-based DOM traits in the subtropical wet (Laojun Mountain, China), temperate arid (Dangjin Mountain, China) and subarctic (Balggesvarri Mountain, Norway) climate zones.The strong relationships between iCER and Gibbs free energy (GFE) were well supported by the results in both field and laboratory experiments shown in Fig. S10.We found the strongest correlations between iCER and almost all traits with two-sided Pearson r values of 0.90 to 0.99 (P < 0.001) in the temperate arid climate zone, which may lead to higher explained variance of the first principal component than the other two climate zones.Abbreviations of DOM traits are detailed in the Methods.

Figure 10 .
Figure 10.The changes in the indicator of compositional-level thermal responses (iCER) against weighted means of formula-based Gibbs free energy (GFE) in the field (a) and laboratory (b) experiments.In the field experiments, we showed their relationships for all nutrients (black line) and each nutrient level (colored lines) in the subtropical wet (Laojun Mountain, China), temperate arid (Dangjin Mountain, China) and subarctic (Balggesvarri Mountain, Norway) climate zones.In the laboratory experiments, we plotted iCER against GFE in the subtropical climate, temperate to subtropical transitional climate, and temperate climate zones.Statistical significance of linear model fits with one-sided F-statistics is indicated by solid (P ≤ 0.05) or dotted (P > 0.05) lines.

Figure S11 .
Figure S11.Density plots show the distribution of compositional-level H/C (a) and O/C (b) ratios of DOM from the field and laboratory experiments in this study, and also from main natural habitats of freshwaters (e.g., lake sediment, lake water, reservoirs, pond, river sediment, river water, and stream water).We synthesized the DOM characteristics derived from FT-ICR MS across the global natural freshwaters that are publicly available until June 2022.More details of global freshwater DOM are shown in Hu et al. (2023) 1 .n is the number of samples in each experiment or natural habitat.

Figure S12 .
Figure S12.The pictures show the dynamics for sediments to settle down after they were

Figure S13 .
Figure S13.Ratios of peak intensities of Run A plotted against Run B for three DOM samples.Ratios of peak intensities were calculated as the ratio of the peak's intensity to the intensity of the base peak at each nominal m/z by following the method in Sleighter et al. (2012) 2 .We compared the relative peak intensities between the replicates (run A and run B) of DOM samples, and found there was a high reproducibility of replicate mass spectra of DOM analyzed by FT-ICR MS.The adjusted R 2 of linear models with one-sided F-statistics was obtained for the relationships between DOM replicates.Dashed line marks 1:1 relationship.

Figure S15 .
Figure S15.Gradual changes in molecular traits (Gibbs free energy, GFE; a) and the percentage of molecular richness (i.e., molecular peak number) for each compound class (b) along the continuum of negative (blue) to positive (red) MERs in the three climate zones.We used the bin size of 2% of all molecules to create the MER continuum, by categorizing all molecules into 50 equal-sized bins according to their magnitude of positive or negative MERs.Blue and red dashed lines in (a) highlight the cutoff of 100% significantly negative and positive MERs for the remaining molecules, respectively.Data in (a) are presented as the means ± s.e averaged across molecules for each removal scenario.