This research document analyzes in detail the long-term impact of tree diversity on carbon stocks and flows in the Sardinilla Tropical Forest Experiment in Panama, part of a global network of experiments with tree diversity (TreeDivNet). International commitments promote large-scale forest restoration as a natural solution to mitigate climate change through carbon sequestration. Growing evidence suggests that mixed forest plantations can sequester more carbon, be less vulnerable to climate extremes, and provide a wider range of ecosystem services than monocultures. However, experimental studies that comprehensively examine the impact of tree diversity on multiple above- and below-ground carbon stocks and flows have been lacking. This study addresses this knowledge gap by using data from The Sardinilla experiment, the oldest experiment with tropical tree diversity, which contains a gradient of single-, two-, three- and five-species mixtures of native tree species. Over 16 years, researchers measured various above- and below-ground carbon stocks and flows, from carbon in above-ground tree biomass, to leaf litter production, to soil organic carbon (SOC).

Methodology

The study was conducted at the Sardinilla site in Panama, on former pasture land dominated by C4 grasses. In 2001, six native tree species were planted: Luehea seemannii, Cordia alliodora, Cashew nut, Hura crepitans, Tabebuia rosea a Cedrela odorata. Species were selected based on their relative growth rate in the natural forests of the region, and fast (Ls, Ca), medium (Ae, Hc) and slow (Tr, Co) growing species were always combined in the mixtures. A total of 24 experimental plots with different diversity were created: 12 monocultures (two for each species), six three-species mixtures and six five-species mixtures. Trees were planted at a constant density of 3 × 3 m. Due to high mortality Ca Only 22 plots were monitored in the first two years after planting over the 16-year experiment. The study analyzed three periods of plantation development: early, middle, and late, which were characterized by repeated climate extremes, including an extremely wet year (2010), a severe drought caused by El Niño (2015), and Hurricane Otto (2016).

Ten components of the forest carbon cycle were measured during three periods of plantation development: carbon in aboveground tree biomass (AGC), carbon in coarse roots (CRC), carbon in coarse dead wood (CWDC), carbon in herbaceous biomass (herbaceousC), carbon production in leaf litter (litterC), soil organic carbon (SOC), leaf litter decomposition, root decomposition, carbon in soil microbial biomass (Cmic) and soil respiration. In addition, canopy opening was measured as a potential cofactor affecting carbon stocks and fluxes. Species-specific and diversity-specific allometric equations were used to estimate aboveground tree biomass (AGB). CRC was estimated based on root to aboveground biomass ratios. CWDC was determined from the weight of fallen branches and trunks. HerbaceousC was measured in quadrats. LitterC was collected in litter traps. SOC and its isotopic fractions (SOC3 and SOC4) were analyzed from soil samples. Leaf and root litter decomposition was measured using nylon bags. Cmic was measured using the substrate-induced respiration method. Soil respiration was measured using a portable infrared gas analyzer. Canopy coverage was measured using hemispherical photographs.

Various statistical methods were used to analyze the data. To compare the components of the carbon cycle after 16 years of growth, multivariate analysis of variance (MANOVA). To investigate temporal changes in carbon stocks and fluxes, mixed-effects ANOVA modelsTo understand the mechanisms behind the impact of tree diversity on carbon dynamics, structural equation models (SEM).

Results

After 16 years of growth, the experimental forest plantation accumulated an average of 35.9 ± 2.7 Mg C ha⁻¹ in trees (AGC + CRC), while SOC decreased by an average of 11.2 ± 1.1 Mg C ha⁻¹, resulting in a net gain of 24.7 ± 2.9 Mg C ha⁻¹. MANOVA revealed significant impact of tree diversity on carbon components directly associated with trees: AGC, CRC and CWDC. This effect was primarily due to AGC, which was 57 % higher in the five-species mixtures (35.7 ± 1.8 Mg C ha⁻¹) than in the monocultures (22.8 ± 3.4 Mg C ha⁻¹). MANOVAs performed with canopy overlap and litterC also showed a significant effect of diversity, with litterC being 64 % higher in the five-species mixtures than in the monocultures. In contrast, none of the MANOVAs performed on soil carbon components revealed a significant effect of tree diversity.

Mixed effects analyses showed that the increase in ΔAGC was significantly slower in monocultures than in most mixtures and the positive effect of tree diversity on ΔAGC tended to strengthen over time in five-species mixtures. CWDC was significantly lower in monocultures than in all mixtures. In contrast, ΔCRC increment did not differ significantly between monocultures and mixtures. HerbaceousC did not differ with diversity but decreased significantly with time. LitterC increased significantly with time and the effect of diversity depended on the period of plantation development, being lowest in monocultures and highest in five-species mixtures at later periods. The predominantly arboreal ΔSOC3 showed significant change with time but not with diversity. ΔSOC4, associated with C4 grasses from previous grazing, varied significantly in response to time and the interaction of time with diversity. Overall, SOC4 decreases were larger than the observed SOC3 increments, resulting in a net negative SOC balance of forest restoration. None of the other variables, including microbial biomass, soil respiration and canopy cover, responded significantly to diversity.

Structural equation models showed that tree diversity significantly reduced canopy cover and increased ΔAGC in both seasons. Canopy cover had a significant positive effect on herbaceousC but reduced litterC. ΔAGC significantly increased litterC and CWDC. These results suggest consistent impact of diversity on aboveground carbon stocks and fluxes and their linkages over timeIn contrast, no significant linkages were observed between aboveground and belowground carbon stocks and fluxes, and only one direct effect of diversity on ΔSOC3 in the late period, which was negative. ΔCRC did not respond significantly to diversity or significantly affect ΔSOC3 in any period.

The results of the study confirm the hypothesis that Tree diversity can increase carbon stocks and fluxes, but this effect was only significant above ground. The observed significant increase in AGC in mixtures with higher diversity is consistent with previous meta-analyses. The long-term carbon balance in restored forests depends not only on the average carbon accumulation, but also on the forest stability and carbon residence time. Although the study does not allow separating the effects of climate extremes from those of stand development, it showed that the positive effects of tree diversity persisted and even strengthened over time, especially for ΔAGC. This suggests greater stability of diverse forest communities against climate fluctuations, which supports the insurance hypothesis.

SEM models revealed that tree diversity directly or indirectly affected all measured aboveground carbon stocks and fluxes, suggesting the interconnectedness of aboveground components of the carbon cycle. Canopy cover played a key role, with denser canopy cover at high diversity increasing litterC but decreasing herbaceousC. Increased tree productivity at high diversity subsequently increased litterC and CWDC, and thus soil carbon fluxes. However, this excess soil carbon flux was largely not incorporated into the soil matrix, and therefore the interconnections of aboveground carbon stocks and fluxes did not significantly extend belowground.

Overall, the Sardinilla forest plantation gained significant above-ground carbon but experienced a loss of SOC in the top 10 cm layer, regardless of tree diversity. This loss of SOC may be associated with the change in land use from pasture to forest and the associated changes in bulk density and SOC concentration. The literature does not provide a clear consensus on the impact of afforestation on SOC. It is possible that the link between tree diversity and SOC will emerge in the Sardinilla plantation over the next decade, highlighting the need for long-term studies.

From a forest restoration perspective, the study results show that Mixed forest plantings can not only increase above-ground carbon stocks and flows compared to monocultures, but also reduce the susceptibility of restored forests to stress and disturbance, thereby increasing carbon stabilityAlthough carbon sequestration is a slow process, natural solutions such as mixed forest plantations are of undeniable importance for carbon sequestration and other ecosystem benefits.

The study showed that Tree diversity significantly increases aboveground carbon stocks and fluxes in the Sardinilla tropical forest experiment, but had no significant effect on belowground carbon stocks and fluxesThe positive effects of diversity on aboveground carbon were strengthened over time, suggesting greater stability of diverse forest communities to climate fluctuations. These findings support the preference for mixed forest plantations over monocultures in forest restoration initiatives aimed at mitigating climate change and emphasize the importance of tree diversity for increasing carbon sequestration potential and ecosystem stability. Spring