Ecological modelling of vascular plant diversity under different climate and land use change scenarios in Nepal Himalaya

Abstract

The effect of climate, topography and land-use on the plant species richness and composition are widely studied topics in ecology. Recent scenarios of climate change studies also show the effect on the species diversity and gain in new area and loss in old habitats. The species richness is considerably high in Nepal with respect to its surface area due to its topography and diversity in micro climates. Climate change and land use change are affecting such richness and composition world widely. Current study was focused on assessing the species richness of vascular plants, their composition and finding the environmental variables which affect them in different elevations and land use gradients in the Nepal Himalaya. We also assessed the effect of climate change on the future distribution of Betula utilis and Larix potaninii var. himalaica which are both tree-line vascular plant species in Nepal Himalaya. The elevation of study sites ranged from 2200 m to 3800 m in six river valleys of Gorkha, Solukhumbu and Taplejung districts of Nepal. Four land use types namely crop field, meadow, exploited forest and natural forest were surveyed. Species were recorded in 25m x 2.5m plots in each land use type. We measured soil temperature as well as air humidity and temperature with data loggers on all study sites. The recorded data show a clear gradient of elevation and land use types designed in current study. Two visits were done for all plots in two different seasons. Although there are some common species recorded in both visits, the species record in individual visits are also unique. The field survey revealed a total of 840 vascular plant species belonging to 492 genera of 120 families. Among them 631 were dicots, 159 monocots, 12 gymnosperms and 38 pteridophytes. There were 96 tree species, 110 shrub species, 586 herbs and 48 climber species. Asteraceae comprised the highest number of species (85 spp.). Multivariate ordination techniques such as Detrended Correspondence Analysis (DCA) and Canonical Correspondence Analysis (CCA) were used to analyze the patterns of species composition and Generalized Linear Model (GLM) was used to find out the significant environmental variables affecting the species richness. In DCA, the first two axes were more than 4.0 standard units inferring high beta diversity. Three sets of environmental variables namely microclimate, bioclimate and topography were used for the ordination. In CCA, isothermality, minimum temperature of the coldest month, precipitation seasonality, precipitation of the coldest quarter were significant contributing bioclimatic variables. Similarly, exposition, land use types, elevation, aspect, slope, region, and valley were vi

significant constraining variables for the species composition. Same set of predictor variables were used to develop three models using GLM. In addition, a synthetic model was developed by combining significant predictor variables from each set of the variables. Linear combination of mean soil temperature, mean humidity, minimum humidity, maximum temperature of warmest month, precipitation of the driest month, precipitation seasonality, valleys, land use types, elevation and latitude produce the best model in forward selection procedure. Maxent modelling was done to predict the future distribution of vascular plants under different climate change scenarios by 2050 and 2070. We chose two species namely Betula utilis and Larix potaninii var. himalaica for this purpose. Both plants occur on high Himalaya up to tree-line. Larix is a deciduous conifer which has limited distribution in central Nepal and adjoining region of Tibet, China. Betula utilis is a deciduous broad-leaved plant with wide distribution from Pakistan to China. It occurs in almost all high mountains of Nepal. The bioclimatic variables were taken from CCSM4 climate database with RCPs 2.6, 4.5, 6.0, 8.5 for 2050 and 2070. Additionally, elevation, aspect and land use types were taken as predictor variables which were taken from the land cover map of Nepal 2010. All models were validated with various statistical tests. All models showed AUC > 0.9, accuracy between 88.4875 and 95.2858, error rate between 0.1151 and 0.0471, and TSS between 0.6359 and 0.7613. The validation parameters showed the robustness of the models. The elevation and annual mean precipitation were main contributing factors for the distribution of Betula utilis and Larix potaninii var. himalaica respectively. The distribution area of Betula utilis will decrease by -18.72% under RCP 2.6 but will increase in all other scenarios by 2050 and 2070. The new gains in area will be distributed towards the western mountains at all RCPs in 2050 and 2070. The predicted area of Larix potaninii var. himalaica did not show specific trend of increment or decrease at all RCPs by 2050 and 2070. The findings of the current study will be applicable in the assessment of effect of environmental and land use variables on the species composition and richness other group of plants too. Current study will be also helpful in devising future policies of conservation of rare, vulnerable, and threatened plant species whose distributions are more likely to be affected by the climate change in the future. Key words: Biodiversity, SDM, Species composition, Species distribution, Treeline