A modelling approach using seedbank and soil properties to predict the relative weed density in organic fields of an Italian pre-alpine valley |
| |
| Authors: | S OTTO M C ZUIN G CHISTȆ & G ZANIN |
| |
| Institution: | National Research Council (CNR), Institute of Agro-environmental and Forest Biology, Legnaro (PD), Italy;, and Istituto Agrario San Michele all'Adige (ISMAA), San Michele all'Adige (TN), Italy |
| |
| Abstract: | In 1996, a study was conducted on the seedbanks of a pre‐alpine valley in northern Italy which had been organically farmed since 1986. The seedbanks were evaluated using soil cores taken from 16 organic fields located at various altitudes and seed numbers were determined using the ‘seedling emergence method’. Thirteen soil properties were also evaluated. In 2003, the germinable seedbank was assessed in five other fields chosen at random. Soil properties were evaluated by the same method as in 1996. Using the data of the first 16 fields as the analysis data set and those of the latter five as an independent validation data set, a quadratic weed seedbank‐soil properties model was built with partial least square regression analysis. The model estimates the relative abundance of the various species as the sum of the contribution of individual soil properties and has a high predictive capacity. With a novel graphic approach, it is possible to describe the nonlinear relationship between each soil property and weed species relative abundance, giving a rational, quantitative, explanation as to why some species are widespread (e.g. Chenopodium album, Galinsoga parviflora and Chenopodium polyspermum), whereas others tend to concentrate in specific fields (e.g. Spergula arvensis). The approach can, when some hypotheses hold, give a rational basis for the explanation of the relative abundance of species in a weed community and constitutes a useful methodology for study and research. |
| |
| Keywords: | seedbank germinable seed soil properties quadratic model partial least square regression analysis |
|
|
