Modeling temporal and spatial effects on plant reproduction using observational data

Abstract

Background and AimsThe persistence of a plant species at local sites depend on species ability to survive and reproduce under local conditions. Estimating environmental influence on reproduction is difficult because climatic and soil conditions have synergistic effects on flower and fruit production, and estimating these synergistic effects require tracking a large number of marked individuals. Here, we develop a statistical method that allows investigating the environmental influence on reproduction using observational data only (no marking required).MethodsWe surveyed 30 standardized permanent plots on central Amazonia for six herbaceous families of Zingiberales and Poales. The plots were surveyed for twelve consecutive months. Using the newly developed method, we estimated the effect of climatic (temporal) and edaphic (spatial) covariates on flower and fruit production.Key resultsWe demonstrate that plant reproduction can be estimated from observational data only when enough temporal and spatial data are available. By using the proposed statistical method, we show that the conversion of flowers into fruits in Amazonian monocots is highest in sandy soils, and the difference in conversion rates between sandy and clayey soils is more pronounced in wetter months.ConclusionsBy comparing the production of flowers and fruits with previously published data on species occurrence, our results suggesting that species distribution is limited primarily by species capacity to produce fruits (e.g. limited pollination) and not by the capacity to produce flowers. Due to the association of fruit production to climatic and edaphic variables, our results point to potential changes in species reproduction and distribution under future climatic regimes.