Modelling actual and potential natural vegetation types: an approach to support the ecological restoration and conservation programmes in Jordan

Abstract

Lying at the junction of Africa, Asia, and the Mediterranean, Jordan is home to a range of diverse and unique vegetation. The country also boasts some of the oldest anthropogenic landscapes in the world due to its location within the Fertile Crescent region. In modern times, however, the area is experiencing significant change due to increasing populations and anthropogenic threats on one hand, and a lack of appropriate planning for the preservation and management of ecosystems on the other. Jordan’s environmental sustainability is also threatened by climate change which increases the vulnerability of many plant species. In the Jordanian context, the application of concepts such as potential, climax, and pristine natural vegetation is difficult as the region has been managed heavily for hundreds of years and mature vegetation is now sparse. Vegetation modelling is desperately needed for agriculture, afforestation, and rangeland management planning and to limit degradation of habitats in Jordan. Most of the existing studies on Jordanian vegetation are insufficient or inadequate in that they often do not take advantage of state-of-the-art satellite imaging and fail to include comprehensive field observations to support mapping of the regions being investigated. Thus, the objective of this study was to model both the current and potential vegetation in Jordan in order to help develop appropriate ecological conservation and restoration programmes in the country. To fulfil this objective, datasets were collected from sampling plots in various habitats and vegetation types, and from satellite images. Then, the data were analysed and interpreted using different state-of-the-art techniques and software. Firstly, hierarchical cluster analysis was used to distinguish vegetation types. This analysis enabled the classification of 16 vegetation types based on the species composition of their perennial vegetation. A reciprocal illumination approach between on-the-ground sampling and satellite imagery was then employed to define the types of vegetation, with datasets obtained from 18 cloud-free Sentinel-2A images. Sentinel remote sensing and GIS software were used to derive a land use/land cover map from high resolution images based on spectral characteristics of the main vegetation types, as verified from the ground data. Prior to field work, an unsupervised map classifying 18 different land use/land cover categories was derived. Based on the updated land use/land cover map, decisions were made regarding where field sampling and ground-based verification was needed. Extensive field experience supported a supervised classification process and the interpretation of satellite images to translate the spectral characteristics into vegetation types. Finally, a vegetation map was produced containing 18 vegetation types. Based on additional information collected during field work, 10 maps were made to illustrate the spatial distribution of human threat to vegetation and its level of impact on habitats. Secondly, species distribution modelling (SDM) was used to predict potential natural vegetation (PNV) in the present-day and the future. Relevant data including indicator species occurrences, climatic data, and topographic data were selected and analysed using presence-only distribution models implemented in the MaxEnt software. This enabled the prediction of the present-day potential distribution of vegetation types. To predict future potential distribution of vegetation, five future climate models were used with three differing carbon emission scenarios for two time periods: 2041-2060 and 2081-2100. Results show a predicted increase in the suitable habitat areas in 2060 and 2100 for some vegetation types: Garrigue and Batha Vegetation, Gravel Hammada Vegetation and Sandy Gravel Hammada Vegetation with Hammada scoparia; Pine and Deciduous Oak Forests in the northwest; Sand Dune Vegetation; and Saline and Thermophilous Vegetation. Conversely, there is a predicted decrease in suitable habitat areas for the same time period for Steppe Vegetation, Juniper and Evergreen Oak Forests, Acacia Woodland, Granite and Sandstone Shrubland, Mudflat Vegetation, Runoff Hammada Vegetation, and Sandy Gravel Hammada Vegetation with Vachellia gerrardii & Artemisia judaica. The ultimate goal of producing these predictive maps was to identify the areas of priority for ecological restoration, review the current boundaries of protected areas, and propose new reserves. The study’s findings are vital for the management, protection, and sustainable utilisation of vegetation in Jordan, with the overall aim of addressing the challenges associated with climate change. Thirdly, results of SDMs were used to identify areas where climate change likely will have little or no impact on vegetation types; these areas are the most appropriate locations for ecological restoration and protection. The established and proposed protected areas network declared by the Government of Jordan were compared with the maps of natural vegetation and potential natural vegetation in order to determine whether the protected areas network will protect the distribution of vegetation types, both in the present-day and in the future. To prioritise conservation areas, a degradation map was used in order to target the most threatened areas. Lastly, the areas that should be targeted by ecological restoration initiatives were identified and used as the basis of a proposal to create eight new protected areas, and to expand the boundaries of eight other existing protected areas.