Reconstructing the range expansion of the European starling in Southern Africa using a hybrid method of niche modelling and individual based modelling

Date
2014-03
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: Much emphasis has been placed on modelling species distributions and less so on the distributional dynamics. It is important to not only ask where species occur, but why and how they reached a specific location. Biological invasions provide an ideal natural experiment for studying the drivers and dynamics of spreading. Here, I examine the drivers and dynamics of the spread of the European Starling, Sturnus vulgaris, in Southern Africa. Since its introduction in Cape Town in 1897, this top invasive avian species has rapidly spread across a large extent of South Africa and is continuously expanding its current range. It is, thus, of great theoretical and management value to elucidate the invasion process and identify key environmental drivers of its range dynamics. The aim of this research is twofold. First, I aim to develop realistic suitable habitat maps for the European Starling in Southern Africa based on multiple climatic and geographic variables using two robust methods of species distribution modelling (SDM), namely maximum entropy (MaxEnt) and boosted regression trees (BRT). Second, I aim to, subsequently, develop a hybrid model that runs a dynamic individual-based model (IBM) on these suitability maps predicted from the SDM. This hybrid model provides an excellent opportunity to incorporate a variety of response regimes on how behavioural strategies and dispersal scenarios respond to environmental and geographic features, and as a result leads to a well-rounded study of the species‟ distribution dynamics in the region. Independent atlas data and field collections were used to parameterize and later validate the model. This allowed me to further identify an optimal model through a parameter sensitivity analysis sorting of the contribution of each environmental and behavioural features to shaping the past and current geographical range of European Starling. According to the optimal model, starlings can choose among five sites to locate the one with the highest habitat suitability. They tend to avoid moving through areas with a 300 m elevation barrier, and the dispersal distance per year is limited to below 200 km. At the regional scale, the starlings were able to manoeuvre around mountainous regions and avoided the semi-desert regions of the Karoo. Their distribution was mainly driven by high winter precipitation along the low-lying coastal regions. Future projections of their distribution suggested a continuous range expansion throughout the provinces of the Free State, Gauteng, North West (reaching Mahikeng, the capital of North West, in 2022) and Limpopo (reaching Polokwane, the capital of Limpopo, in 2046) as well as into the neighbouring countries of Mozambique and Zimbabwe (reaching Bulawayo, 2nd largest city in Zimbabwe, in 2062).
AFRIKAANSE OPSOMMING: Klem word dikwels meer geplaas op die modellering van spesies se verspreiding en minder op die dinamika agter die verspreiding. Dit is belangrik om nie slegs te vra waar spesies voorkom nie, maar hoekom en hoe hulle by 'n spesifieke plek beland. Biologiese indringing bied 'n ideale natuurlike eksperiment vir die bestudering van die bestuurders en dinamika agter spesies se verspreiding. Hier ondersoek ek die bestuurders en dinamika agter die verspreiding van die Europese Spreeu, Sturnus vulgaris, in Suider-Afrika. Sedert sy bekendstelling in Kaapstad in 1897, het hierdie top indringer voël vinnig versprei oor 'n groot gebied van Suid-Afrika en is tans steeds besig om verder te versprei. Dit is dus van groot teoretiese en beskermings waarde om die indringers proses te verstaan en die noodsaaklikste omgewings bestuurders van die spesies se verspreidings dinamika te identifiseer. Die doel van hierdie navorsing is tweeledig. Eerstens, streef ek daarna om realistiese geskikte habitat kaarte vir die Europese Spreeu in Suider-Afrika te ontwikkel wat gebaseer is op verskeie klimaat en geografiese veranderlikes met behulp van twee robuuste metodes van verspreiding spesies modellering (species distribution modelling, SDM), naamlik maksimum entropie (MaxEnt) en regressie bome (boosted regression trees, BRT). Tweedens, streef ek daarna om 'n hibriede model te ontwikkel wat 'n dinamiese individu-gebaseerde model (IBM) op hierdie geskiktheids kaarte, voorspel deur die SDM, te simuleer. Die model bied 'n uitstekende geleentheid om 'n verskeidenheid van reaksie kombinasies te integreer oor hoe gedrags-strategieë en verspreiding scenario's reageer op omgewing en geografiese kenmerke, en as 'n gevolg lei tot 'n afgeronde studie van die spesie se verspreiding dinamika in die streek. Onafhanklike atlas data en veld data is gebruik om die model te parameteriseer en die geldigheid daarvan te toets. Hierdeur kon ons „n optimale model identifiseer deur middel van 'n parameter sensitiwiteits-analise sortering van die bydrae van elke omgewings en gedrags kenmerk tot die vorming van die verlede en die huidige geografiese verspreiding van die Europese Spreeu. Volgens die optimale model, kon die spreeus tussen vyf plekke die een met die hoogste habitat geskiktheid kies. Hulle is geneig om gebiede met 'n 300 m hoogte versperring te vermy, en die verspreidings afstand per jaar is beperk tot minder as 200 km. Op die plaaslike skaal, die spreeus was in staat daartoe om te beweeg rondom bergagtige gebiede en die semi-woestyn gebiede van die Karoo te vermy. Hul verspreiding is hoofsaaklik gedryf deur 'n hoë winter reënval langs die laagliggende kusgebiede. Toekomstige projeksies van hul verspreiding stel voor dat verdere verspreiding plaasvind in die provinsies van die Vrystaat, Gauteng, Noord-Wes (waar hulle Mahikeng, die hoofstad van Noord-Wes, bereik in 2022) en Limpopo (Polokwane, die hoofstad van Limpopo, word bereik in 2046) sowel as in die buurt lande van Mosambiek en Zimbabwe (Bulawayo, 2 grootste stad in Zimbabwe, word bereik in 2062).
Description
Thesis (MSc)--Stellenbosch University, 2014.
Keywords
Biogeography, Biological invations -- Mathematical models, European Starling -- Geographical distribution -- Southern Africa, Sturnus vulgaris -- Dispersal behavior, Biological invasions -- Environmental aspects, European starling -- Dispersal behavior, UCTD
Citation