Vegetation dynamics in a semi-arid freshwater wetland: a measure for restoration success

Abstract

The functioning of wetland ecosystems is impacted by human anthropogenic activities, which affect wetlands’ vegetation development and their ability to offer ecosystem services. The need to understand vegetations’ natural regeneration, in previously degraded wetlands of semi-arid regions is essential to measure restoration success. Nylsvley freshwater wetland is the largest inland flood plain found in the semi-arid region of South Africa and a Ramsar site of worldwide conservation importance. It provides a habitat to some endangered species; however, it is facing degradation problems leading to a declining of its biodiversity. The importance of Nylsvley wetland prompted an initiation into its ecological restoration, to try and reverse the ecosystem degradation and monitor vegetation development of this Ramsar site. This study aimed to evaluate restoration efforts, through an analysis of vegetation dynamics, in a restored semi-arid freshwater wetland, so as to understand the success story of restoration efforts. The first objective of the study examined historical rainfall trends in the Nyl River system as rainfall affects vegetation productivity. Daily rainfall data was measured from eight rainfall stations for the period, 1950 to 2016, to generate seasonal and annual rainfall data. The Mann-Kendall and quantile regression were applied to assess trends in this rainfall data. Normalized Difference Vegetation Index derived from satellite images from between 1984 and 2003, utilising zonal statistics was correlated with rainfall of the same period to assess vegetation dynamics. The Mann-Kendall and Sen’s slope estimator showed that only one station had significant increased rainfall trend, annually and seasonally, at p < 0.05, whereas all the other stations showed insignificant trends during these rainfall seasons. Quantile regression showed that 50% and 62.5% of the stations had increasing annual and seasonal rainfall, respectively. Three of the eight stations had data which were statistically significant at p < 0.05, indicating increasing and decreasing rainfall trends. These rainfall trends show that the rainfall at Nylsvley decreased Vegetation dynamics in a semi-arid freshwater wetland: a measure for restoration success between 1995 and 2003. The R2 between rainfall and Normalized Difference Vegetation Index of Nylsvley is 55% indicating the influence of rainfall variability on vegetation productivity. The second objective assessed spatio-temporal changes of natural vegetation development within Vogelfontein, part of Nylsvley, utilising 1984, 2005 and 2016 satellite imagery which provided at least 10 years’ evaluation of conditions, before and after restoration. Land cover maps were generated from classified Landsat images using maximum likelihood classification algorithm with five land uses - old fields, water, bare ground, sparse and dense vegetation. The results showed a worsening of natural vegetation regeneration from 1984 until 2005, as bare ground and old field increased by over 19%. This decline was attributed to cultivation, which ended in 1996 in Vogelfontein. The period 2005 to 2016 saw the classes gaining in natural recovery of vegetation (7%). This indicates the necessity for wetland restoration, as there was evidence of vegetation recovery in a landscape facing a degradation trajectory. Overall, the area covered by vegetation from the 1984 to 2016 period remained in the negative, despite the increase of vegetation cover between the 2005 to 2016 period. The third objective assessed vegetation development in three project sites, namely, A61C-01, A61C-02 and A61C-03, to evaluate restoration interventions made and how they are linked to vegetation development in Nylsvley. Initial monitoring results showed restored sites with monitoring data at implementation stage with identified impacts of hydrology, geomorphology and vegetation and their change score, based on the WET-health assessment recorded. The photographs from previous documents played a key role in providing wetland status before and during implementation in addition to WET-Health data. The records indicated that to improve flood attenuation, road strip with reno gabion structure was constructed across the wetland; to curb soil loss, berms and road strips have been constructed on most roads and some degraded roads were closed to allow vegetation regeneration. The natural hydrological function of the wetland was improved by destroying berms previously constructed to divert water from the floodplain. Vegetation dynamics in a semi-arid freshwater wetland: a measure for restoration success Local plant biodiversity was improved by fencing off previously-degraded areas from grazers, whilst old trees along the berms were kept, as they are potential seed banks. Finally, the fourth objective aimed to quantify composition and diversity of species in Nylsvley wetland to comprehensively outline the vegetation restoration success. Ten circular quadrants of 10-meter radius were utilised using Adapted Point Centre Quarter method to collect woody vegetation structural measurements. The Simpson’s Diversity Index was used to determine diversity and composition of species in the measured sites. The Analysis of Variance results shows that reference plots have significantly more trees, higher species richness, as well as taller and more woody species density per hectare than the treatment plots. This study provides beneficial information for management to understand the influence of historical rainfall patterns on climate variability in the Nyl River system and how they impact on vegetation growth and recovery of the reserve as there is a strong relationship between rainfall and Normalized Difference Vegetation Index of Nylsvley. The study will also help researchers to understand the annual and seasonal variability of rainfall over the study region and should serve as a foundation for further studies. The study will further shape the wetland restoration discourse through evaluation of ecosystems restoration efforts that have not been widely adopted. The study utilizes Geographic Information Systems and Remote Sensing tools as management strategies to monitor vegetation-cover change, before and after restoration in Nylsvley. This extends the frontiers of knowledge through providing an understanding of natural changes in vegetation, in space and time; this is a critical indicator of vegetation development and restoration success of Nylsvley. Finally, an evaluation of the restoration activities done in Nylsvley and the quantification of vegetation species in the restored sites will form the basis for additional monitoring of vegetation regeneration in Nylsvley and with a potential for upscaling to other similar semi-arid freshwater wetlands.