Drivers of weed diversity on patch and pattern scale in vineyards of Kosovo

Abstract

The PhD thesis provides data on the vineyard weed flora and vegetation in Kosovo, especially in the well-known wine-growing regions Rahovec and Suhareka. The data collected and analyzed are important in two ways: On the one hand, they contribute to reducing knowledge gaps on biodiversity and its drivers in Kosovo. Furthermore, the study can help regional agriculture to implement weed management strategies adapted to the species and vegetation that are present. The investigation methods include (A) field surveys of the weed flora and vegetation and (B) soil chemical analysis in cultivated fields of the study areas, (C) germination experiments in climate chambers, and (D) germination experiments under field conditions. In the germination experiments, four weed species occurring in the two study regions were investigated. The field survey and the field experiments were conducted in Kosovo, the germination experiments in climate chambers at the University of Giessen. In May to August 2016 and 2017, the weed flora and vegetation was documented on a total number of 70 vineyard patches (35 patches in Rahovec; 35 patches in Suhareka) (A). On each of the 70 patches, four plots were selected randomly for detailed vegetation sampling. Each of the plots was divided into two subplots ‘within’ and ‘between rows of vines’ with a standard size of 1 m x 5 m. To avoid edge effects, the minimum distance of each subplot to the patch border was 5 m. For each subplot vegetation relevés were compiled. For each subplot, additional site data was documented in the field and derived from maps (height above sea level, exposition, aspect, northness, eastness). Based on interviews with the farmers who cultivated the patches, additional information was compiled for each of the 70 patches on herbicide use. This made it possible to distinguish between two management types (with versus without herbicide use) in the data analysis. On each of the subplots soil sampling (soil depth: 5-15 cm) was carried out (B). In each subplot, the soil of three sub-samples that were taken with a hand auger was mixed. Each of the mixed soil samples was dried at room temperature and sieved (to exclude stones with a size >2 mm from further soil analysis and to quantify the stone content of the soils) before the chemical analysis. Soil pH was measured in both H2Odest. and CaCl2. The total amount of organic matter (OM) was determined by the ignition method. The amount of plant-available P and K was quantified with the help of the Ammonium lactate method. The vegetation and land management data and the soil data were considered in ordinations and indicator species analysis to relate the diversity of the weed vegetation to the considered management practices and site conditions. For four selected weed species (Amaranthus retroflexus, Chenopodium album, Echinochloa crus-galli, Solanum nigrum) relations between air temperature and germination were analyzed in climate chambers (C). The seeds were sampled on the patches that were investigated in 2016. The germination experiment was conducted with seeds of the F1-generation of the sampled seeds. The germination succeed was tested for different temperatures ranging from 3 °C to 30 °C. Five replicates with a standard number of 50 seeds per species and temperature level were analyzed. The germination experiment lasted for 12 weeks in winter 2017/18. In addition to the climate chamber experiment, relations between air temperature and germination were studied under field conditions (D). As described before, seeds of the F1-generation of the sampled seeds were also used in the field experiment in 2018. In standardized plots (30 cm x 30 cm) with heat-sterilized soil (to prevent germination of seeds that may occur in the soil) a standardized number of 200 seeds of the selected weed species was investigated by means of weekly recordings within the entire vegetation period. In these recordings the number of germinated seeds was documented. For each of the selected species, three replicates were studied to allow statistical analysis of the data. To relate the germination data to the air temperature during the vegetation period the minimal and maximal air temperature next to the plots was recorded daily. Overall 84 weed species were recorded during 2016 and 2017 in the two study regions. In both years, the species number was higher in Suhareca region. The species numbers and weed coverage differed considerably between the two regions and management types. Fields without herbicide application in Suhareka region had significantly more species than herbicide-treated ones in Rahovec region. The DCA analysis and the indicator species analysis prove remarkable differences in the weed composition and coverage between the study regions, management types and the small-scale differences within and between the rows of vines in 2016 and 2017. Also the indicator species analysis resulted in differences between the two study regions, management practices and the situation within or between rows of vines. The mean pH value in H2Odest. and the mean amount of OM were higher in Suhareka compared to Rahovec region, and the mean amount of OM was higher within compared to between rows in both regions. Moreover, the DCA analysis showed a correlation between the vegetation and soil chemical parameters. The germination experiments in climate chambers and in the field resulted in species specific germination behavior and in differences in the germination rates between the two study regions. There were significant differences between regions at the different temperatures regarding germination of F1 seed of Amaranthus retroflexus and Chenopodium album. Moreover, Echinochloa crus-galli, Chenopodium album and Solanum nigrum achieved higher germination rates from seeds collected in the Suhareka region. The summarized results can be interpreted in several ways: The current diversity of the vegetation in the studied vineyards is generally low due to implemented weed control measures. However, At the patch scale, the land-use intensity (with versus without herbicide application; lower intensity ‘within rows’ than ‘between rows’) is a driver of vegetation diversity, while at the pattern scale, differences in land-use intensity between the two study regions and also differences in soil chemical characteristics affect vegetation diversity. The region-dependent differences in temperature-dependent germination rates found in the germination experiments with the F1 generation of selected weed species indicate genetic differences between the populations studied. Since both regions are not far from each other within Kosovo, these differences could be a consequence of different intensities of cultivation in both regions and resulting genetic differences of the populations. The results are discussed with recommendations for a more sustainable use of the wine growing areas of Kosovo in the future. Options for organic management and future soil management with plant cover are presented in regard to their importance for ecosystem services and biodiversity protection.