Naslov
Analysing impact of spatial accuracy and sample size on habitat suitability models for invasive plants

Autori
Jelaska, Sven D. ; Orešković, Anja ; Nikolić, Toni

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
Vegetation patterns in natural and cultural landscapes - Abstract books / Guarino, Riccardo ; Bazan, Giuseppe ; Barbera, Giuseppe - Palermo : Palermo University Press, 2017, 182-182

ISBN
978-88-99934-43-9

Skup
60th IABS Annual Symposium

Mjesto i datum
Palermo, Italija, 20.06.2017. - 24.06.2017

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
modelling ; Croatia ; maps

Sažetak
Although the survey of invasive plants in Croatia has intensified in the last 10 years, there hasn’t been a systematic survey of their distributions, and available data differ with respect to their spatial accuracy. All chorological data in Flora Croatica Database (FCD) have assigned levels of spatial precision on a scale from 1 (least precise) to 11 (coordinates obtained with GPS). We have used this information to classify data and test how the precision affects the environmental preferences (seasonal mean temperature and precipitation ; slope ; aspect ; distance from settlements and roads and human population density) of each species and to what extent spatial accuracy of chorological data, as well as sample size, influence the outcome of habitat suitability models. Using Maxent models with a conservative threshold value of 0.5, we reclassified output grids to create maps of suitable habitats for invasive plants based on different sets of data with respect to their spatial accuracy and sample size. When we increased the size of samples by adding less precise data, statistical significant positive correlation was observed between sample size and area designated as suitable habitat. This wasn’t the case when we compared standalone data sets with different level of spatial precision. The same trend was observed for invasive and native plants that share similar environmental envelopes (e.g. Erigeron anuus vs. Arrhenatherum elatius and Leucanthemum vulgare ; Robinia pseudoacacia vs. Carpinus betulus and Quercus petraea). Furthermore, when comparing maps of suitable habitats derived by cumulative addition of data with lower spatial precision, the increase of area occurred in a “nested manner” i.e. spreading from the centre that was predicted with the best data. Contrary, when data with different spatial precision was used as standalone, predicted areas did not differ, but their spatial position did. Comparison of different sample size, within same spatial precision categories, did not yield significant differences, neither in predicted area or spatial arrangement. In summary, the spatial precision of data proved to be more critical than sample size for predicting the distribution of invasive species, suggesting that future mapping efforts should try to be spatially as precise as possible.

Izvorni jezik
Engleski

Znanstvena područja
Biologija

POVEZANOST RADA