Orsolya Valkó

Classification of Herbaceous Vegetation Using Airborne Hyperspectral Imagery

Alkali landscapes hold an extremely fine-scale mosaic of several vegetation types, thus it seems challenging to separate these classes by remote sensing. Our aim was to test the applicability of different image classification methods of hyperspectral data in this complex situation. To reach the highest classification accuracy, we tested traditional image classifiers (maximum likelihood classifier—MLC), machine learning algorithms (support vector machine—SVM, random forest—RF) and feature extraction (minimum noise fraction (MNF)-transformation) on training datasets of different sizes. Digital images were acquired from an AISA EAGLE II hyperspectral sensor of 128 contiguous bands (400–1000 nm), a spectral sampling of 5 nm bandwidth and a ground pixel size of 1 m. For the classification, we established twenty vegetation classes based on the dominant species, canopy height, and total vegetation cover. Image classification was applied to the original and MNF (minimum noise fraction) transformed dataset with various training sample sizes between 10 and 30 pixels. In order to select the optimal number of the transformed features, we applied SVM, RF and MLC classification to 2–15 MNF transformed bands. In the case of the original bands, SVM and RF classifiers provided high accuracy irrespective of the number of the training pixels. We found that SVM and RF produced the best accuracy when using the first nine MNF transformed bands; involving further features did not increase classification accuracy. SVM and RF provided high accuracies with the transformed bands, especially in the case of the aggregated groups. Even MLC provided high accuracy with 30 training pixels (80.78%), but the use of a smaller training dataset (10 training pixels) significantly reduced the accuracy of classification (52.56%). Our results suggest that in alkali landscapes, the application of SVM is a feasible solution, as it provided the highest accuracies compared to RF and MLC. SVM was not sensitive in the training sample size, which makes it an adequate tool when only a limited number of training pixels are available for some classes.

Traditional cattle grazing in a mosaic alkali landscape: effects on grassland biodiversity along a moisture gradient.

Extensively managed pastures are of crucial importance in sustaining biodiversity both in local- and landscape-level. Thus, re-introduction of traditional grazing management is a crucial issue in grassland conservation actions worldwide. Traditional grazing with robust cattle breeds in low stocking rates is considered to be especially useful to mimic natural grazing regimes, but well documented case-studies are surprisingly rare on this topic. Our goal was to evaluate the effectiveness of traditional Hungarian Grey cattle grazing as a conservation action in a mosaic alkali landscape. We asked the following questions: (i) How does cattle grazing affect species composition and diversity of the grasslands? (ii) What are the effects of grazing on short-lived and perennial noxious species? (iii) Are there distinct effects of grazing in dry-, mesophilous- and wet grassland types? Vegetation of fenced and grazed plots in a 200-ha sized habitat complex (secondary dry grasslands and pristine mesophilous- and wet alkali grasslands) was sampled from 2006–2009 in East-Hungary. We found higher diversity scores in grazed plots compared to fenced ones in mesophilous- and wet grasslands. Higher cover of noxious species was typical in fenced plots compared to their grazed counterparts in the last year in every studied grassland type. We found an increasing effect of grazing from the dry- towards the wet grassland types. The year-to-year differences also followed similar pattern: the site-dependent effects were the lowest in the dry grassland and an increasing effect was detected along the moisture gradient. We found that extensive Hungarian Grey cattle grazing is an effective tool to suppress noxious species and to create a mosaic vegetation structure, which enables to maintain high species richness in the landscape. Hungarian Grey cattle can feed in open habitats along long moisture gradient, thus in highly mosaic landscapes this breed can be the most suitable livestock type.

Litter and graminoid biomass accumulation suppresses weedy forbs in grassland restoration

Abstract Accumulated biomass of sown species and litter plays an important role in success of restoration projects. We studied the effects of litter and graminoid biomass on species richness and biomass of early colonising forbs in former alfalfa fields sown with seed mixtures containing seeds of native grass species (Festuca pseudovina, Festuca rupicola, Poa angustifolia, Bromus inermis, 2005). The amount of litter, forbs and graminoids was measured in the first 3 years after sowing (2006–2008). Ten aboveground biomass samples (20 cm × 20 cm) per field were collected in June every year. We found significantly lower forb biomass in the second and third year, than in the first year after sowing. Litter and biomass of graminoids increased significantly during the study, and correlated negatively with the biomass and species richness of forbs. Mean scores of litter and graminoid biomass were two to three times higher in sown fields than in native grasslands. Our results suggest that the accumulation of litter and graminoid biomass is beneficial in suppression of weedy forbs, but in the long run it might also hamper the immigration of target species.

Solonetz meadow vegetation (Beckmannion eruciformis) in East-Hungary – an alliance driven by moisture and salinity

Wet meadows are transitional habitats between wetlands and dry-mesophilous grasslands; thus, they are vital in sustaining biodiversity as sources, stepping stones and green corridors of species dispersal. It is especially valid for inland alkali vegetation, where high landscape-scale patchiness is typical and rapid shifts between vegetation types occur frequently. Solonetz meadows (Beckmannion eruciformis) are among the best-preserved open habitats in Europe harbouring a unique flora including several endemics. Besides their importance, studies on the vegetation of solonetz meadows are scarce even though this knowledge would be vital for their effective conservation and management. Using own records and literature data, we provide a synthesis of the alliance Beckmannion eruciformis to charac-terise its associations floristically and ecologically in five regions along the river Tisza, East Hungary. We studied three associations of the alliance: (i) Agrostio stoloniferae-Alopecuretum pratensis, (ii) Agrostio stoloniferae-Beckmannietum eruciformis and (iii) Agrostio stoloniferae-Glycerietum pedi-cellatae. We found that solonetz meadow associations were separated along a moisture gradient with Agrostio stoloniferae-Alopecuretum pratensis at the drier end and Agrostio stoloniferae-Glycerietum pedicellateae at the wet end. This gradient was also justified by the distribution of the phytosociological groups. The proportion of species of Festuco-Brometea division was the highest in Agrostio stolonifer-ae-Alopecuretum pratensis, while the proportion of Cypero-Phragmitetea species was the highest in Agrostio stoloniferae-Glycerietum pedicellateae. Species of Puccinellio-Salicornetea had the highest proportion in Agrostio stoloniferae-Beckmannietum eruciformis, indicating the high soil salinity of this association. Our results suggest that Agrostio stoloniferae-Alopecuretum pratensis association plays an important role in preserving the continental flora elements of the surrounding dry grasslands. The ecological indicator values for soil moisture and salinity suggest that in case of the studied solonetz meadow associations, humidity increases with decreasing elevation, while salinity is highest at medium elevations. Our results suggest that Agrostio stoloniferae-Beckmannietum eruciformis is a transition towards salt marsh associations, while Agrostio stoloniferae-Glycerietum pedicellatae is a transition towards freshwater marshes. Our results show that preserving solonetz meadows is an important task for nature conservation as they have a crucial role in maintaining landscape-scale species and habitat diversity and act as buffer zones around wetlands.

Supporting biodiversity by prescribed burning in grasslands — A multi-taxa approach

There are contrasting opinions on the use of prescribed burning management in European grasslands. On the one hand, prescribed burning can be effectively used for the management of open landscapes, controlling dominant species, reducing accumulated litter or decreasing wildfire risk. On the other hand burning can have a detrimental impact on grassland biodiversity by supporting competitor grasses and by threatening several rare and endangered species, especially arthropods. We studied the effects of prescribed burning in alkaline grasslands of high conservation interest. Our aim was to test whether dormant-season prescribed burning can be an alternative conservation measure in these grasslands. We selected six sites in East-Hungary: in three sites, a prescribed fire was applied in November 2011, while three sites remained unburnt. We studied the effects of burning on soil characteristics, plant biomass and on the composition of vegetation and arthropod assemblages (isopods, spiders, ground beetles and rove beetles). Soil pH, organic matter, potassium and phosphorous did not change, but soluble salt content increased significantly in the burnt sites. Prescribed burning had several positive effects from the nature conservation viewpoint. Shannon diversity and the number of flowering shoots were higher, and the cover of the dominant grass Festuca pseudovina was lower in the burnt sites. Graminoid biomass was lower, while total, green and forb biomass were higher in the burnt plots compared to the control. The key finding of our study was that prescribed burning did not decrease the abundance and diversity of arthropod taxa. Species-level analyses showed that out of the most abundant invertebrate species, 10 were not affected, 1 was negatively and 1 was positively affected by burning. Moreover, our results suggest that prescribed burning leaving unburnt patches can be a viable management tool in open landscapes, because it supports plant diversity and does not threaten arthropods.

Rainfall fluctuations and vegetation patterns in alkali grasslands - using self-organizing maps to visualise vegetation dynamics

Knowledge about the drivers of vegetation dynamics in grasslands is fundamental to select appro-priate management for conservation purposes. In this study, we provide a detailed analysis of vegeta-tion dynamics in alkali grasslands, a priority habitat of the Natura 2000 network. We studied vegetation dynamics in five stands of four alkali grassland types in the Hortobagy National Park (eastern Hunga-ry), between 2009 and 2011. We analysed the effect of fluctuations in precipitation on both the overall vegetation composition and on the cover of each species using Self Organizing Map neural networks (SOM). We found that SOM is a promising tool to reveal plant community dynamics. As we analysed species cover and overall vegetation composition separately, we were able to identify the species re-sponsible for particular vegetation changes. Fluctuations in precipitation (a dry season, followed by a wet and an average season) caused quick shifts in plant species composition because of an increasing cover of halophyte forbs, probably because of salinisation. We observed a similar effect of stress from waterlogging in all studied grassland types. The species composition of Puccinellia grasslands was the most stable over the three years with varying precipitation. This was important as this grassland type contained many threatened halophyte species. Self-organising maps revealed small-scale vegetation changes and provided a detailed visualisation of short-term vegetation dynamics, thus we suggest that the application of this method is also promising to reveal community dynamics in more species-rich habitat types or landscapes.

Abandonment of croplands: problem or chance for grassland restoration? case studies from hungary

Abstract In Central‐ and Eastern Europe, the collapse of socialist regimes resulted in a transformation of state‐owned agricultural cooperatives to privately owned lands from the early 1990s onwards. These socioeconomic processes resulted in landscape‐scale changes in biodiversity, ecosystem services and agricultural production. In parallel, large‐scale abandonment of croplands, especially on sandy, salty or frequently inundated areas, became common. Abandoned croplands are usually sensitive to species invasions, and are hotspots of noxious weeds, posing threats both to agriculture and nature conservation. Grassland restoration on former croplands can be an effective strategy for suppressing these species. Thus, a common goal of nature conservation and agriculture can be the restoration of grasslands on former croplands to (1) suppress weed and/or invasive species in line with the policy “Good Farming Practices”, (2) support animal husbandry by creating meadows or pastures, and to (3) recover biodiversity and ecosystem services. In the present paper we report “best practices” of grassland restoration projects from Hungary. Our aim was to compare the effectiveness of spontaneous grassland recovery vs. active grassland restoration by seed sowing in terms of the recovery of biodiversity and ecosystem services, such as weed control and biomass production. Our results showed that grassland restoration on abandoned fields offers a viable solution for restoring biodiversity and ecosystem services. Seed sowing ensures higher weed control and biomass production, but results in lower biodiversity compared to spontaneous recovery. Both restoration methods can be cost‐effective, or even profitable even within a relatively short period of a nature conservation project.

New measurements of thousand-seed weights of species in the Pannonian flora

For understanding local and regional seed dispersal and plant establishment processes and for considering the ecotypes and other forms of specific variability, hard data of locally or regionally measured traits are necessary. We provided newly measured seed weight data of 193 taxa, out of which 24 taxa had not been represented in the SID, LEDA or BiolFlor databases. Our new measurements and formerly published data of locally collected seed weight records together covers over 70% of the Pannonian flora. However, there is still a considerable lack in seed weight data of taxonomically problematic genera, even though they are represented in the Pannonian flora with a relatively high number of species and/or subspecies (e.g. Sorbus, Rosa, Rubus, Crataegus and Hieracium). Our regional database contains very sporadic data on aquatic plants (including also numerous invasive species reported from Hungary and neighbouring countries) and some rare weeds distributed in the southwestern part of the country. These fact...

Early vegetation development after grassland restoration by sowing low-diversity seed mixtures in former sunflower and cereal fields

We studied the early vegetation dynamics in former croplands (sunflower and cereal fields) sown with a low-diversity seed mixture (composed of 2 native grass species) in Egyek-Pusztakócs, Hortobágy National Park, East-Hungary. The percentage cover of vascular plants was recorded in 4 permanent plots per field on 7 restored fields between 2006 and 2009. Ten aboveground biomass samples per field were also collected in June in each year. We addressed two questions: (i) How do seed sowing and annual mowing affect the species richness, biomass and cover of weeds? (ii) How fast does the cover of sown grasses develop after seed sowing? Weedy species were characteristic in the first year after sowing. In the second and third year their cover and species richness decreased. From the second year onwards the cover of perennial grasses increased. Spontaneously immigrating species characteristic to the reference grasslands were also detected with low cover scores. Short-lived weeds were suppressed as their cover and biomass significantly decreased during the study. The amount of litter and sown grass biomass increased progressively. However, perennial weed cover, especially the cover of Cirsium arvense increased substantially. Our results suggest that grassland vegetation can be recovered by sowing low diversity mixtures followed up by yearly mowing. Suppression of perennial weed cover needs more frequent mowing (multiple times a year) or grazing.

The effects of micro-habitats and grazing intensity on the vegetation of burial mounds in the Kazakh steppes

Background: Burial mounds (kurgans) of Eurasian steppes are man-made habitat islands that have the potential to harbour rich plant diversity due to micro-habitats associated with their topography. Aims: We assessed whether kurgan micro-habitats harboured different species pools and functional groups from those found on the surrounding steppes. In addition, we asked if these mounds were affected by different grazing intensities from those on the surrounding vegetation. Methods: We surveyed kurgan micro-habitats (northern and southern slopes, surrounding ditch) and adjacent steppe plains in non-grazed, moderately grazed and heavily grazed sites in northern Kazakhstan. We analysed differences in species composition of four habitats under three grazing regimes using Generalised Linear Mixed Models, PCA ordination and indicator species analysis. Results: Kurgan micro-habitats had diverse vegetation and supported the co-existence of plant species with different environmental needs. We identified 16 steppe specialists confined to kurgan micro-habitats. Steppe vegetation was well-adapted to extensive grazing, although heavy grazing supported ruderals and a decline in steppe specialists. There was a significant interaction between grazing intensity and habitat type: heavy grazing supported ruderals and suppressed steppe specialists especially on the slopes. Conclusions: We highlighted that kurgans play an important role as maintaining high plant diversity locally in extensive steppe plains in Central-Asia by increasing environmental heterogeneity and supporting specialist species confined to these micro-habitats.