Functional-trait based restoration alters nutrient cycling and invasion rates in Hawaiian lowland wet forest

Many degraded ecosystems have altered nutrient dynamics, due to invaders possessing a suite of traits that allow them to both outcompete native species and alter the environment. In ecosystems where invasive species have increased nutrient turnover rates, it can be difficult to reduce nutrient availability. This study examined whether a functional-trait-based restoration approach, involving the planting of species with conservative nutrient-use traits, can slow rates of nutrient cycling and consequently reduce rates of invasion. We examined a functional trait restoration initiative in a heavily-invaded lowland wet forest site in Hilo, Hawai'i. Native and introduced species were chosen to create four experimental hybrid forest communities, in comparison to the invaded forest, with a factorial design in which communities varied in rates of carbon turnover (slow or moderate [SLOW, MOD]), and in the relationship of species in trait space (redundant or complementary [RED, COMP]). After the first five years, we evaluated community-level outcomes related to nutrient cycling: carbon (C), nitrogen (N), and phosphorus (P) via litterfall, litter decomposition, and also outplant productivity and rates of invasion. We found that: (1) regardless of treatment, the experimental communities had low rates of nutrient cycling through litterfall relative to the invaded reference forest, (2) the MOD communities had greater nutrient release via litterfall than the SLOW communities, (3) introduced species had greater nutrient release than native species in the two MOD experimental communities, and (4) within treatments, there was a positive relationship between nutrient levels and outplant basal area, but outplant basal area was negatively associated with rates of invasion. The negative relationships among basal area and weed invasion, particularly for the two COMP treatments, suggest species existing in different parts of trait space may help confer some degree of invasion resistance. The diversification of trait space was facilitated by the use of introduced species; a new concept in Hawaiian forest management. Although challenges remain in endeavors to restore this heavily degraded ecosystem, this study provides evidence that functional trait-based restoration approaches using carefully crafted hybrid communities can reduce rates of nutrient cycling and invasion in order to reach management goals.Excel Funding provided by: National Science FoundationCrossref Funder Registry ID: http://dx.doi.org/10.13039/100000001Award Number: 1754844Funding provided by: Strategic Environmental Research and Development ProgramCrossref Funder Registry ID: http://dx.doi.org/10.13039/100013316Award Number: RC-2117Funding provided by: Hawai'i Army National Guard Environmental Office*Crossref Funder Registry ID: Award Number:Nutrient cycling metrics Litterfall was measured as an index of productivity. Litter screens (40 cm x 40 cm, 1 mm mesh size) were placed in the plots along diagonal and crossed transect lines (n=20 per plot). Litter was collected monthly, bulked per plot, dried at 70°C for at least 48 hours, and then weighed. Starting in 2016, we separated out litter from the outplant species and sorted and weighed each species in order to calculate species-specific productivity and overall outplant contribution to the plots. Nutrients in the litterfall of outplant species were determined by grinding material in a Wiley mill (40 mesh), and then processing samples with a Costech CE Elemental Analyzer (Costech ECS 4010; Valencia, CA) to obtain the concentration of carbon (C) and nitrogen (N), and a Varian inductively coupled plasma spectrometer (Varian Vista-MPX inductively coupled plasma-optical emission spectrometry; Palo Alto, CA) to obtain the concentration of P. All nutrient analyses were conducted at the University of Hawai′i at Hilo Analytical Laboratory. We also examined the decomposition rate of each outplanted species using litterfall collected in 2016. Litter bags with the dimensions of 20 cm x 30 cm (area inside) were constructed of 1-mm mesh fiberglass window screen, filled with 5 grams of litter each. Bags were placed in another area of the forest used for past experiments (Ostertag et al. 2009; Kandert et al. 2021). Bags (n = 10 per time period per species) were collected at intervals of 4 and 12 months. Non-litter items such as roots and small rocks were discarded prior to drying (70°C for 48 hours) and weighing. Samples were ashed at 500°C for five hours (Muffle furnace Thermolyne 1400; University of Hawai′i at Hilo Analytical Laboratory) to determine the percentage of organic material and mass was corrected accordingly. The proportion of mass loss was calculated at each time point. Nutrient release for each species was determined with a series of two calculations that combined litterfall rates, litter nutrient concentrations, and decomposition rates per species. The series of equations used were: 1. Litterfall mass (g/plot) * (C, N, or P (%)/100) = C, N or P in litterfall mass (g C, N, or P/plot) 2. C, N or P in litterfall mass (g C, N, or P/plot) * (Mass Loss (%)/100) = C, N or P released from litterfall (g C, N, or P/plot) These equations provided the nutrient release per species per plot and for treatment comparisons, we added up values from all 10 species per plot. Productivity & invasive resistance metrics In May 2014 all outplants were tagged and their locations within plots mapped using ArcPad on Allegro MX field computers. All stems ≥ 1.3 m height that had a diameter at breast height (DBH) of at least 1.0 cm were measured to calculate basal area at the individual- and then plot-level. All existing natives and outplants were censused annually to monitor changes in survival and growth. Although the number of outplant individuals with a DBH ≥ 1.0 cm was minimal in 2014, 25% of outplants were recorded with a DBH by 2017. In using basal area/plot as a proxy for productivity it should be noted that this metric is not inclusive for all outplants but rather the most productive as gauged by those reaching the 1.0 cm DBH threshold. From the moment plots were originally cleared, it was necessary to hand-weed at approximately 3-4-mo intervals to sustain outplant growth. We removed all undesired recruits (any seedling or resprout that was not native or of a species outplanted in the experiment). After two years of plant community development, we were able to reduce weeding frequency to 6-mo intervals. At the start of 2016 and the years following, we began recording the number of weed species found in the plots, as well as the number of person-hours taken to weed a plot. These data points quantify invasion resistance in the analyses