Characteristics of leaf nutrient resorption efficiency in Tibetan alpine permafrost ecosystems

Abstract

Nutrient resorption is an important strategy for nutrient conservation, especially in permafrost ecosystems where plant growth is limited by nutrients. Based on the measurements mainly derived from tropical, subtropical and temperate regions, current projections suggest that resorption efficiency is higher for leaf nitrogen (N) than for phosphorus (P) in cold regions. However, these projections have not been fully validated due to the lack of observations in permafrost ecosystems. Here, we carry out a large-scale sampling campaign along a permafrost transect on the Tibetan Plateau. Our results show that, in contrast with the prevailing view, resorption efficiency is higher for leaf P than N in permafrost ecosystems (75.1 ± 1.8% vs. 58.7 ± 1.5%; mean ± standard error). Our results also reveal that leaf P resorption efficiency is higher in permafrost ecosystems than in global herbaceous plants, while there is no difference for leaf N resorption efficiency. Interestingly, there is a trade-off between leaf N resorption efficiency and soil N mineralization rate, but no such pattern exists for P. These results illustrate the unique characteristics of plant nutrient resorption in permafrost ecosystems and advance our understanding of nutrient conservation strategies in little-studied permafrost regions.

Fig. 1. Geographic distributions of study sites.

a The distribution of non-herbs (gray dots) and herbs (orange dots) in the global database. The global database was compiled from published datasets (see details in Supplementary Note 4). b The layout of sampling sites across the Tibetan alpine permafrost region. The map was created using ArcMap 10.7 (Environmental Systems Research Institute, Inc., Redlands, CA, USA) based on data derived from the National Snow & Ice Data Center (https://nsidc.org/data/ggd318/versions/2)/(https://nsidc.org/about/data-use-and-copyright) and Zou et al. (https://tc.copernicus.org/articles/11/2527/2017/)/CC BY (https://creativecommons.org/licenses/by/3.0/), respectively.

Fig. 2. Concentration-based leaf nutrient resorption efficiencies in plants among Tibetan alpine grasslands, global forbs, and graminoids.

a Assessing the differences in leaf N and P resorption efficiencies in plants across Tibetan alpine grasslands with a paired-samples t-test (two-sided, n = 30). b Comparative analysis of leaf nutrient resorption efficiencies between plants across Tibetan alpine grasslands and global forbs and graminoids with independent-samples t-tests (two-sided). The global database was compiled from published studies (see details in Supplementary Note 4). Sample sizes of leaf N and P resorption efficiencies are 83 and 65 for graminoids, 43 and 27 for forbs in global datasets, respectively. Data are represented as the means ± SE (standard error). *** (unadjusted P < 0.001), and different letters (unadjusted P < 0.05) represent significant differences (lowercase letters for N and capital letters for P).

Fig. 3. Assessment of three control strategies underlying plant nutrient resorption across Tibetan alpine grasslands.

a Relationship between N concentrations in the mature and senescent leaves. b Association between P concentrations in the mature and senescent leaves. c Linkage between resorbed N:P and leaf N:P concentrations in the mature leaves. d Correlation between log₁₀-transformed resorbed N:P and log₁₀-transformed leaf N:P concentrations in the mature leaves. The gray dashed line in each panel is the 1:1 line. Only significant relationships are shown using solid lines. Error bars denote SE of mean at each site (n = 3). Statistics (slope, R², and P value) are shown for the linear mixed-effects models with two-sided t-tests (unadjusted P < 0.05).

Fig. 4. Trade-off between plant nutrient resorption and soil nutrient supply across Tibetan alpine grasslands.

a Relationship of leaf N resorption efficiency with topsoil N mineralization rate. b Association of leaf P resorption efficiency with topsoil P mineralization rate. A significant relationship is shown by a solid line. Error bars denote SE of mean at each site (n = 3). Statistics (R² and P value) are shown for the generalized linear mixed-effects models with two-sided t-tests (unadjusted P < 0.05).