Increases and decreases in soil moisture in water-limited plant communities cause asymmetrical responses in biomass but not in diversity

Abstract

Aims: Changes in precipitation patterns, such as the predicted increases in the frequency of climatic extremes, are likely to alter plant communities, but whether responses to drought or to wetter conditions respectively cause consistent, opposite responses is debated. Here, we assessed the response in biomass production and species diversity of water-limited plant communities to the direction (increase or decrease) and the magnitude (micro- and macro-climatic effects) of changes in soil moisture.

Location and methods: We reciprocally translocated soils containing seedbanks from two climates (semi-arid and mediterranean) at a micro-climatic (opposite slopes) and a macro-climatic scale (between climates) in Chile.

Results: The biomass production for the soils that were translocated from wetter to drier climates was unrelated to the available soil moisture. The lowest biomass was produced in the wettest climate on the wet slope. Biomass production increased after a translocation to the drier climate (representing the largest change in climate). Nonetheless, the highest overall biomass for the wet to dry translocation was produced on the mediterranean dry slope with intermediate soil moisture. However, on the same mediterranean dry slope, biomass was almost zero for the soil translocated the other way around (i.e. drier to wetter). Diversity after 24 months was unaffected by micro-climatic change, but soils transplanted towards the drier climate yielded a plant community with increased diversity.

Conclusion: Our results showed direction and magnitude of climate change but also the response factor which is studied matters to detect direction-dependent responses, i.e. species richness had a linear and reversible response; however, the response of biomass depended on the origin of the transplanted material (soil and plant community), indicating history dependence, i.e. hysteresis. This emphasizes that responses to unidirectional climate manipulation experiments may not be able to capture the entire nature of the response of plant communities to climate change.

Methods

The dataset is related to the publication “Increases and decreases in soil moisture in water-limited plant communities cause asymmetrical responses in biomass but not in diversity”, by van den Brink et al. 2024 in Journal of Vegetation Science. The study assessed the response in biomass production and species diversity of water-limited plant communities to the direction (increase or decrease) and the magnitude (micro- and macro-climatic effects) of changes in soil moisture.

Thus, the dataset comprises plant biomass, species richness and diversity data from a translocation experiment between two study sites (semi-arid, Mediterranean), and within them between opposite slopes. 

The Excel file contains two sheets, the first one with the important metadata, and the second one contains the full data.

General methods

We reciprocally translocated soil with seed banks between the semi-arid and mediterranean climate (Figure 1a), to understand the response of plant communities to large (macro-) climatic changes. Additionally, we used the local variation of the landscape within the sites to assess the response of these communities to small (micro-) climatic changes by reciprocally transplanting soil with seeds between dry (north-facing) and wet (south-facing) slopes.

On each slope, one plot (50 X 50 cm), with four subplots (of 20 x 20 x 5 cm each) was established (Figure 1b). In each of the four subplots, the top five cm of the soil (including the organic layer) was excavated during the late dry season (March-April 2016) and mixed to obtain the best possible homogeneous distribution of soil and seeds per plot. The homogenized soil was split into four equal parts and each part was placed in a paper bag. One part was taken to the other climate (500 km away) and placed in a randomly assigned subplot of the randomly assigned “partner”-plot with the same slope (“macro-climatic change”, Figure 1b). Two other parts were also taken on a 500 km trip but returned to the original site. The trip served as a procedural control to ensure equal treatment of the transplanted material between climates and within sites (including a control) as soil aggregates might disintegrate during the translocation and influence the results (Amézketa, 1999). From these two parts, one was returned to the original plot in a randomly assigned subplot (“control”). The second part was placed in a randomly assigned subplot of a randomly assigned “partner”-plot on the opposite slope within the same climate (“micro-climatic change”). We included the methodological hypothesis that biomass production and species diversity could be influenced by travel. For this, we immediately placed a fourth part back into a randomly assigned subplot (“control without travel”). All soils were returned to the field before the first winter rain.

The transplanted soil was monitored for two years, and all the species that germinated were annuals. The germinated individuals were therefore considered to have survived if they flowered and were then clipped at ground level shortly before seed dispersal to avoid contamination of local plant communities with foreign seeds. Their dry biomass was measured (after at least 48h of drying at 40°C), and they were identified to calculate the species richness and diversity. Because the first year was very dry, there was very little seed production, and thus, during the second year, no seedlings from surrounding species were found in the soil that was translocated between climates, indicating that our plots contained only the seeds that came from the original soil. After two years, the “non-local” soil and all remaining roots were carefully removed and disposed of.

Total biomass production (g of dry weight) was calculated as the sum of the biomass of all species per subplot. Species diversity was measured using the Shannon-Wiener diversity index (H; sensu Wilhm, 1968) at the subplot level. To calculate the Shannon-Wiener Index (H), we used the species biomass rather than the number of individuals: H = -Σ[(pi) x ln(pi)], where pi is the proportion of total biomass represented by species i.

Code/Software

We used ANOVAs with origin (semi-arid vs. mediterranean), original slope (dry vs. wet), treatment (control, micro-climatic change and macro-climatic change) and all their interactions as fixed factors. Although the experiment was full-factorial, only a subset of the data addressed our ecological questions. Therefore, we focused the analysis on three comparisons/contrasts (Kwon, 1996) that addressed our hypotheses, comparing between origins and among treatments, and using the data from the three-way interactions (origin*original slope*treatment) (see Casper and Castelli, 2007 for another example on planned contrasts): 1) we compared the controls, to analyze the natural differences between the climates and slopes; 2) we then compared the translocations from dry-to-wet (semi-arid dry slope vs. semi-arid wet slope vs. mediterranean dry slope, see Figure 1), and from wet-to-dry (mediterranean wet slope vs. mediterranean dry slope vs. semi-arid wet slope, Figure 1), 3) we additionally compared the reciprocal translocations for micro-climatic changes (between slopes in each climate) in order to understand the influence of micro- and macro-climatic changes, in two directions, on biomass production and species diversity after two years. P-values were adjusted with the Benjamini-Hochberg procedure using a false discovery rate of 0.1. We chose 0.1 because we have a maximum of 6 contrasts, reducing the false discovery rate to almost 0. All statistical analyses were performed using JMP 15.

Subject keywords

Biomass, Climate change, Species diversity, Biological sciences, extreme climatic events, Mediterranean plant communities, plant productivity, precipitation change, reciprocal transplant, soil transplant, species richness