Urban ecosystems correspond to the classes at first and second levels, as defined in MAES (Maes et al. 2013) and include urban, industrial, commercial and transport areas, urban green areas, mines, dumping and man-made sites. At the third level, the typology of urban ecosystems in Bulgaria corresponds to the National concept for spatial development for the period 2013-2025. Different types of urban ecosystems in Bulgaria are defined in Table 1.

The urban ecosystems in Karlovo are formed by seven types Fig. 1.

Figure 1.  

Spatial distribution of urban ecosystems in Karlovo (Urban ecosystem subtypes codes are explained in Table 1).

In this study, four areas of the urban territory of Karlovo were investigated: the area of ‘Apostolova gora’ park; the central part of the city; the ‘Hunting’ Park and the city stadium area; the area of ‘Suchurum’ waterfall (Northern Park). The selected areas include the most important elements of the town’s green infrastructure (Fig. 2).

Figure 2.  

Studied areas in Karlovo municipality.

Two in-situ observations in the selected case-study sites were conducted during the periods 25-27 May and 7-8 July 2016 and a control observation of some individual trees in the central part of city was carried out on 9 August 2017. The health status of the vegetation was evaluated on ten individuals of each species (except in cases where the number of host plants was smaller) (Sinclair et al. 1987). The harmful effects and impact of arthropod pests (insects, mites) and fungal pathogens causing damage to green ecosystems were evaluated on the basis of the percentage of crown defoliation and discolouration of leaves or needles following the methodology of the International Co-operative Programme ‘Forests’ (Eichhorn et al. 2010). Defoliation as well as discolouration of crowns were estimated visually as the main parameters of the crown condition vitality. The estimated values were given in 5 classes from 0% (healthy tree) to 100% (dead tree), subdivided into five classes (Table 2).

Defoliation surveys were linked with detailed assessments of biotic and abiotic damage causes on leaves, stems, branches and young shoots. Other specific indicators were also recorded: presence of drying or rotted trees, endangering human life and health; insect attacks causing allergies; invasive species penetration; presence of trees colonised by highly aggressive xylophages or dangerous pathogens being a threat to the surrounding plants.

Determination of damage symptoms and causes (biological agents or abiotic factors) were implemented for studying of cause-and-effect mechanisms. For the purpose of this study, three categories of harmful effects caused by pests were determined according to the gradation of native species, detection of invasive alien species, range of trophic groups that pests belong to and the physiological state of host plants (Table 3). The reason for such categorisation is the fact that defoliations caused by phylophages could be overcome relatively easily at the expense of the reserve capacities. To their difference, the damage on non-renewable and difficult-to-renew tissues and organs leads to a decline in the permanent health state and death of the infested trees (Georgiev et al. 2017). Concerning fungal pathogens causing damage to tree and shrub vegetation, three categories of harmful effects on the basis of pathogen virulence and aggression, physiological condition of host plants and opportunities for wide spreading of disease, were also suggested.

Considering the role of the UGI for providing ES, the authors assessed the ‘quality’ of GI in selected case-study areas in Karlovo by a multispectral camera ‘Parrot SEQUOIA’. The camera was integrated with a specialised professional UAV system of type ‘Flying Wing’, which provides the opportunity to collect specialised information about the state of GI in urban areas in an extremely effective way. The multispectral camera provides imaging of the studied area in 4 channels of the electromagnetic spectrum: G (Green corresponds to the reflected energy in the 530–570 nm spectral band), R (Red- reflected energy in the 640–680 nm spectral band), Red Edge (very narrow band between 730–740 nm, corresponding to the rapid change from low Red reflectance to high Near Infrared reflectance) and NIR (Near infrared- wavelengths in the 770 nm to 810 nm range). It has also a standard RGB channel. The camera was equipped with a solar radiation sensor which serves for calibration of the obtained reflex images.

For accomplishment of the assessment, NDVI (Normalised Difference Vegetation Index) was used, which was obtained by digital mixing of imagery, captured in the red and near-infrared (NIR) range and was calculated according to the following formula:

\(NDVI=(NIR-Red)/(NIR+Red)\)

NDVI is a simple graphical indicator used to assess whether the target being observed contains live green vegetation or not. The index is very helpful in the examination of the vitality of plants by capturing the amount of light which is being absorbed and reflected by them and the available chlorophyl l. NDVI itself varies between -1.0 and +1.0 and typically the green vegetation has values between 0.3 and 1. In general, if there is much more reflected radiation in near-infrared wavelengths than in visible wavelengths, then the vegetation in that pixel is likely to be dense and may contain some type of forest. Subsequent work has shown that the NDVI is directly related to the photosynthetic capacity and hence energy absorption of plant canopies (Myneni et al. 1995). Visualisation of the imaged area in a different colour range shows the vitality level of plants and clarifies where the plants are healthy and where additional care is needed.

The field mission was preceded by careful planning and drawing up of a precise flight plan. The missions’ planning was performed using the SenseFly-E-motion specialised platform. Flights were planned in four selected areas of the urban territory of Karlovo (Fig. 2).

Imaging was performed with sensors for gathering of positioning and quality data (about the condition of the ecosystems) with the assistance of the professional UAV system eBee that utilises photogrammetric and multispectral cameras. The gathered field information was processed by a specialised platform for photogrammetric processing Pix4D Professional Mapper.

The remote sensing was combined with an in-situ observation of the health status of tree vegetation and analyses of plant diversity.

As a result of the in-situ observations and assessment of the urban vegetation condition in selected areas in Karlovo, it was found that the city is characterised by a well-developed green infrastructure with a prevalence of native tree species. A total of 41 tree and shrub species were assessed for changes in crown defoliation and discolouration as a parameter of tree vitality. Twenty-eight of the species were deciduous and the remaining thirteen were conifers. The mean defoliation for all observed species ranges between 0% and 100% (Table 2). In good condition, with no symptoms of damage on any assessed sites, were trees of species Ailanthus altissima, Abies alba, Cercis siliquastrum, Diospyros lotus, Fraxinus ornus, Tilia argentea, T. platyphillos, Sambucus nigra and Sophora japonica. Amongst the rest of the observed species, 32% were assessed as slightly defoliated, 48% were with moderate mean defoliation and 20% were severely defoliated (mean defoliation between 70-100%). Furthermore, it was observed that trees, evaluated as moderate and severe degrees, revealed damage caused by a number of different biotic, abiotic and anthropogenic factors. However, the assessment of health condition in different parks varied depending on site and tree specific conditions. Amongst the common groups of trees, the most severely affected were species from genera Pinus and Ulmus. Identified insects, diseases and other damaging agents that adversely affected the ability of green ecosystems of Karlovo to produce goods and services, were subdivided into three categories based on their harmful effects.

Eighteen insect species (Insecta) and one mite species (Arachnida) were established on forest trees and shrub species in urban systems of Karlovo (see Table 4). Eight of them belong to the first harmful category: four invasive species (Cameraria ohridella, Cydalima perspectalis, Gilletteella cooleyi and Corythucha ciliata), three destructive xylophagous species (Zeuzera pyrina, Phaenops cyanea and Scolytus multistriatus) and one species causing allergies (Thaumetopoea pityocampa). The second harmful category includes five species (Pissodes sp., Phyllonorycter platani, Gossyparia spuria, Monochamus galloprovincialis pistor and Psylla buxi) and the third category – six species (Tetraneura ulmi, Aegosoma scabricorne, Cerambix cerdo, C. scopolii, Rhagium inquisitor and Eriophyes tristriatus).

Twenty-five pathogen species were identified as causal agents of biotic damage (see Table 4). Amongst them, seven species belong to first harmful category - six strongly pathogenic and invasive species (Allantophomopsis pseudotsugae, Botryosphaeria dothidea, Cryptostroma corticale, Diplodia sapinea, Dothistroma pini and Ophiostoma novo-ulmi) and two fungi endangering human health and life (C. corticale and Phellinus igniarius). The rest of the identified pathogens were divided into a second harmful category because of their local influence on host plants' leaves or needles. Saprophyte fungi, belonging to the third harmful category, were not observed on dead plants.

Amongst conifers, pine trees displayed the highest mean defoliation in ‘Apostolova gora’ and central parks. In ‘Apostolova gora’ park, all pine trees showed severe defoliation (50-90%) due to a combined attack by pests and pathogens belonging to the first harmful category. Insect pests were the predominant identified cause of this damage. The most dangerous species was the buprestid Phaenops cyanea on Pinus nigra and P. radiata (Fig. 3). Fungal pathogens were the second major causal agent group affecting all assessed pine trees causing needle and shoot blight diseases. The invasive pathogenic fungi Dothistroma pini and Diplodia sapinea affected P. nigra, P. radiata and P. strobus and caused an unattractive environment and degradation of ecosystems in the urban landscape. The third major identified cause of damage to pine trees was abiotic and anthropogenic agents. Drought periods within the observed two growing seasons additionally led to worsening of the tree physiology in the second year in such a way that they became more vulnerable to pest attack which caused severe damage and death to trees. In 2016, the mean defoliation of P. radiata trees in the central part of the city was assessed between 40 and 70%. The main causal agents and factors responsible for the observed damage symptoms were the pests P. cyanea, Pissodes sp., M. galloprovincialis pistor, R. inquisitor and development of the invasive strongly pathogenic fungi D. pini and D. sapinea (first harmful category). In 2017, a control observation of P. radiata trees revealed that almost half of thetrees were completely dead and the rest were dying with defoliation 70-90%, heavily affected by the coincident development of the pests and pathogens (Fig. 4).

Figure 3.  

Damage on Pinus nigra and P. radiata caused by Phaenops cyanea: A – larval galleries (Apostolova gora park, 26 May 2016); B – exit holes of adults (Central park, 9 August 2017); C – weakened and dead trees (Central park, 9 August 2017).

Figure 4.  

Damage on Ulmus minor caused by Ophiostoma novo-ulmi in Central park: A – dying tree (8 July 2016); B – dead tree (9 August 2017); C – galleries of elm bark beetle, Scolytus multistriatus (9 August 2017).

All observed elm trees (Ulmus minor) in the central and ‘Hunting’ parks were attacked by the invasive pathogen Ophiostoma novo-ulmi which caused one of the most destructive diseases – Dutch elm disease, distributing along vascular tissues, which significantly disturbed the balance of the tree water and caused tracheomycotic damage, severe wilting and tree mortality (first harmful category). During the first observation in the central parks (May 2016), only slight symptoms of the disease were recorded. Two months later (July 2016), the proportion of declined elm trees, as well as their defoliation, significantly increased (Fig. 4). The crowns of infected trees were covered by withering leaves which remained on the branches. The results obtained in the control survey (August 2017) revealed that all trees infected by the fungus in the previous year were completely dead and the population density of the pathogen’s vector (S. multistriatus) was extremely high (Table 4).

Concering established dangerous phylophagous insects from the first harmful category (C. ohridella, C. perspectalis, G. cooleyi, C. ciliata and T. pityocampa), the impact on observed hosts was not significant due to their low density.

Two introduced invasive fungal pathogens were detectednd these caused significant damage and a negative influence on tree vitality - Botryosphaeria dothidea detected on Sequoia sempervirens and Juniperus virginiana in the observed central parks and Cryptostroma corticale observed on drying and completely dead trees of Acer pseudoplatanus in the central and ‘Hunting’ parks.

Development of timber-destructive fungi on tree stems and branches endangering human life and health (first harmful category) was observed on Salix babylonica trees along the ‘Stara reka’ river in the area of ‘Suchurum’ waterfall park.

Fungal pathogens, belonging to the second harmful category and causing local, non-aesthetic damage to green vegetation used for planting, were determined in all selected case-study areas. On the leaves of horse chestnut (A. hippocastanum), the invasive species Phyllosticta paviae had developed together with the leaf mining moth (C. ohridella). Pathogens from genera Lophodermium were detected on needles of Pinus spp. and Picea abies and caused slight defoliation of crowns.

Mechanical damage made by persons working or visiting the parks was observed on single trees in all selected sites. Wounding of trees initiated the process and led to their decay and death. During the in-situ observation conducted on 25 May 2016, very strong winds passed through the city and broke off brittle branches of Acer dasycarpum in ‘Suchurum’ waterfall park, resulting in threats for human injury or death.

The territory of ‘Apostolova gora’ park has been captured with 1008 individual images, covering 62.9 ha with an average resolution of 13.2 cm. A view of the 4 captured channels and the generated Digital Surface Model is shown in Fig. 5. After processing of the basе products (images and Digital Surface Model), the NDVI model for the territory was created (Fig. 6).

Figure 5.  

Digital Surface Model of ‘Apostolova gora’ park.

Figure 6.  

NDVI model of ‘Apostolova gora’ park (Min- minimal NDVI score for the class, max-maximal NDVI score for the class, area (ha) total area of the NDVI class, area (%)- relative share of the NDVI class within the studied area).

The map clearly shows the areas with a high value of vitality of vegetation cover, as well as those with a lack of vegetation cover and vegetation in an unsatisfactory condition. In almost 50% of the studied area, the vegetation index is more than 0.75, but in the park territory, relatively large areas can be seen with poor vegetation and the corresponding NDVI index values lie between 0.6 and 0.7.

The territory of the central part of the city has been captured with 1004 individual images, covering 61.2 ha with an average resolution of 13.2 cm. A view of the 4 captured channels and the generated Digital Surface Model is shown in Fig. 7. After processing of the basе products (images and Digital Surface Model), the NDVI model for the examined territory was created (Fig. 8).

Figure 7.  

Digital Surface Model of the central part of Karlovo.

Figure 8.  

NDVI model of the Central park of Karlovo (Min- minimal NDVI score for the class, max-maximal NDVI score for the class, area (ha) total area of the NDVI class, area (%)- relative share of the NDVI class within the studied area).

Generally, in this part of the urban area of Karlovo, the elements of green infrastructure are the most fragmented and have the smallest coverage. More than 50% of the territory represents embedded and built-up areas without vegetation cover. Only about 15% of the central part of the city has an NDVI index more than 0.75.

The ‘Hunting’ park and the city stadium area have been captured with 852 individual images, covering 33.6 ha with an average resolution of 13.75 cm (Fig. 9). After the processing of the basе products (images and Digital Surface Model), the NDVI model was created (Fig. 10).

Figure 9.  

Digital Surface Model of ‘Hunting’ park and city stadium area of Karlovo.

Figure 10.  

NDVI model of ‘Hunting’ park and city stadium area of Karlovo (Min- minimal NDVI score for the class, max-maximal NDVI score for the class, area (ha) total area of the NDVI class, area (%)- relative share of the NDVI class within the studied area).

Around 20% of the examined territory has NDVI indices above 75% which is due to the large areas with green infrastructure and good vitality of the vegetation. Data analysis shows that, in the ‘Hunting’ park, there are areas with a low level of plant vitality, corresponding to a low level of chlorophyll content.

The area of ‘Suchurum’ waterfall in Karlovo has been captured with 860 individual images, covering 51.2 ha with an average resolution of 15.1 cm (Fig. 11). After processing of the basе products (images and Digital Surface Model), the type of NDVI model for the examined territory was created (Fig. 12).

Figure 11.  

Digital Surface Model of the area of ‘Suchurum’ waterfall in Karlovo.

Figure 12.  

NDVI model of the area of ‘Suchurum’ waterfall in Karlovo (Min- minimal NDVI score for the class, max-maximal NDVI score for the class, area (ha) total area of the NDVI class, area (%)- relative share of the NDVI class within the studied area).

The examined area is characterised by the relatively good condition of the green system. For more than 40% of the whole territory, the index values above 0.7 were calculated. These values show a high level of plant vitality and relatively good spatial coverage.