Using expert and spectral methods to assess visually attractive urban informal green spaces in Lublin, Poland.

doi:10.21203/rs.3.rs-2187110/v2

Download PDF

Research Article

Using expert and spectral methods to assess visually attractive urban informal green spaces in Lublin, Poland.

https://doi.org/10.21203/rs.3.rs-2187110/v2

This work is licensed under a CC BY 4.0 License

Version 2

posted

You are reading this latest preprint version

The objectives of the study is to develop the criteria for expert assesment of visual attractiveness of informal urban green spaces and compare results with indicators derived from spectral indexes and GIS data. The study area is Lublin in Poland, a medium-sized European city. Expert assessment includes: general assessment - defining overall attractiveness, naturalness, landscape contrast and uniqueness. We juxtaposed the results with spectral indexes and GIS data. The analyses were performed for the target areas and a 300 m buffers around them. The spectral indexes largely confirm the results of the expert assessment. Clear relationship exists between overall attractiveness and temperature and between landscape contrast and temperature. Areas with greater landscape contrast also have a larger difference in index values between the area and the buffer. There is a positive correlation between overall attractiveness and land slope. Spectral indexes also indicated the value of some sites that received a lower expert assessment. We suggest that the most visually attractive areas be called “enchanted natural places" (ENPs) as a tool for formalising the protection of these areas. Spectral indexes confirm the values of ENPs, but also shown the value of sites with lower expert assessment, more common “bushes and shrubs”.

perceiving the landscape

urban wilderness

spectral indexes

remote sensing

In times of new challenges for urban landscapes, many authors present different ways of broadly approaching the issue of urban greenery, understanding green spaces (Taylor and Hochuli 2017), formulating their typology, such as UGS – urban green spaces (Cvejić, Eler, Pintar, Železnikar, Haase, Kabisch and Strohbach 2015) and measuring their values. Informal, semi-natural, wild areas occupy an important place among them. These areas are addressed by Rupprecht and Byrne (2014), Rupprecht, Byrne, Ueda and Lo (2015), who refer to them as “informal green spaces” (IGS) and define them as “any urban space with a history of strong anthropogenic disturbance that is covered at least partly with non-remnant, spontaneous vegetation”. The scholars also distinguish IGS from formal green spaces and areas overgrown with remnant vegetation, including protected areas. This approach is close to the theory of four kinds of nature (Kowarik 2005, 2018), in which IGS are categorised in group 4 – novel urban ecosystems. On the other hand, Sikorska, Łaszkiewicz, Krauze and Sikorski (2020) postulate the extension of IGS to remnant areas, including protected areas, arguing that they are characterised by a similar low degree of human interference, and coins the phrase “unmanaged informal green spaces” to name them. We also use this term in our study. Research on informal green spaces are conducted with different approaches and methods. Among others researchers focuses on visual attractiveness and perception of these areas and research using spectral data and GIS databases. Broader research on the indicators of visual attractiveness of landscapes was made by Tveit, Ode and Fry (2006) and Ode, Fry, Tveit, Messager and Miller (2009), as well, the beauty indicators were developed by Kerebel, Gélinas, Déry, Voigt and Munson (2019). Many researchers also study the perception and value of these areas for residents (Rink and Emmrich 2005; Kendal, Ford, Anderson and Farrar 2015; Rupprecht et al. 2015; Pietrzyk-Kaszyńska, Czepkiewicz and Kronenberg 2017; Sikorska et al. 2020; Włodarczyk-Marciniak, Sikorska, Krauze 2020). Research of informal green spaces by remote sensing data and GIS databases also gives important results for delineation and assessment of the values. These reasearch were conducted by Carver, Comber, McMorran and Nutter (2011), Müller, Bøcher and Svenning (2015), Yang, He, Wang, Yan, Yu, Bu, Yang, Chang and Zhang (2017), Müller, Bøcher, Fischer and Svenning (2018), Zheng, Chen, Pan, Cai and Chen (2019), Cheng, Guan, Zhou, Zhao and Zhou (2019), Zhu, Wang, Yan, Liu and Zhou (2019), Sun, Tan, Chen, Song, Zhu and Hou (2020), Ossola, Jenerette, McGrath, Chow, Hughes and Leishman (2021), Shah, Garg and Mishra (2021) and Hauser, Féret, An Binh, van der Windt, Sil, Timmermans, Soudzilovskaia and van Bodegom (2021). Despite the variety of approaches, definitions and classifications, most authors point to the important role of these areas and the need for their further exploration. Kremer et al. (2016) emphasise the need to combine different research approaches on ecosystem services, including urban green areas. Following these authors, we juxtaposed the two approaches and investigated the relationship occurring between visual assessment and indices obtained from remote sensing (spectral) data for unmanaged informal green spaces in the city. The aim of the research was to develop an expert assessment of unmanaged IGS in terms of their visual appeal, and find out how to distinguish them from other unmanaged IGS; to identify the most landscape-attractive unmanaged informal green spaces in the city, which we called “enchanted natural places” (ENPs), uroczyska miejskie in Polish and to compare the results of the expert assessment with the data obtained from spectral indexes and GIS data of these areas and their immediate surroundings, for relationships.

2.1. Study area

The study area was Lublin, a medium-sized Central-European city, and the most populous in eastern Poland. The city occupies 147 km², with a population of 321,324 inhabitants. The city is situated at 163-238 m above sea level, within the Lublin Upland (Statistics Poland 2018). The Bystrzyca River and its three tributaries – the Czechówka, the Czerniejówka and the Nędznica – and two periodic watercourses flow through the city. Two distinct parts of Lublin can be distinguished. The western part is characterised by a thick loess cover, intersected by numerous dry valleys and gorges, which creates a diverse landform. The majority of residential areas, usually on hilltops, are located in this part. In land depressions there are green areas, as well as communication routes. The eastern part, on the other hand, is devoid of a loess cover and has a less diversified, rolling terrain. In this part industrial development and smaller clusters of residential development predominate. The highly diverse landform features and watercourses have an impact on remarkable flora and landscape diversity. Research on this topic was carried out from 2016. Detailed field inventory was made between 2016 and 2018 in different seasons and weather conditions. Of all urban green spaces (UGS), we focused on the “unmanaged informal green spaces” after Sikorska et al. (2020). They include IGS (Rupprecht and Byrne 2014), sites of remnant vegetation (Florgård 2007) and protected areas.

2.2. GIS and remote sensing data

In our study, we used satellite imagery from the Operational Land Imager (OLI) sensor on the Landsat 8 satellite. Data was obtained from the Earth Explorer service of the USGS (U.S. Geological Survey) (EarthExplorer 2021). The image was recorded on 14 August 2018, 09:25 am, in 11 spectral channels from the visible and near-infrared (VNIR) and short-wave infrared (SWIR) (430 - 2290 nm) and thermal infrared (TIR) (1060 - 1251 nm) ranges. The spatial resolution of the image was 30 m and the radiometric resolution was 12 bits (USGS 2021). For the study, we used VNIR (channels 2, 3, 4, 5) and SWIR (channels 6, 7) channels and channel 10 of the TIR range. A calibration procedure was carried out for the image: transformation of Radiance values to Surface Reflectance (VNIR, SWIR) and conversion of channel 10's DN (Digital Number) row value to the Land Surface Temperature (LST) value on the day the image was taken (Fig. 1 and 2). For this purpose we used the ATCOR module with Catalyst Professional software (Catalyst.Earth 2021). We also used cartographic data, orthophotos and the Digital Elevation Model (DEM), which were acquired free of charge from the Polish Central Office of Geodesy and Cartography (www.geoportal.gov.pl). DEMs were acquired in the ARC/INFO ASCII GRID format, containing the height value of points in a regular 1-metre grid. The data was processed into an image and then the slope was calculated from it (Zevenbergen and Thorne 1987). The Slope for the assessment of the land surface topography was used. The Urban Atlas 2018 database (UA2018), Copernicus programme was used too. The UA2018 database contains pan-European land cover and land use data for functional urban areas (Copernicus 2021). For the purpose of our analysis, we combined the following categories: Construction sites; Continuous urban fabric; Discontinuous dense urban fabric; Discontinuous low density urban fabric; Discontinuous medium density urban fabric; Discontinuous very low density urban fabric; Industrial, commercial, public, military and private units; Fast transit roads and associated land; Other roads and associated land; Isolated structures; Mineral extraction and dump sites. The table with full data: expert assessment, spectral indexes and digital data for all sites is added in supplementary materials (Appendix A).

2.3. Course of the research

Based on cartographic analyses and available databases (geoportal.gov.pl), we identified 91 unmanaged IGS with expected visual attractive landscape (Carver et al. 2012; Müller et al. 2018; Ode et al. 2008; Tveit et al. 2006). We also analysed the surrounding areas. For this purpose, we set a buffer of 300 m for each area as a zone of closest impact between the study area and its surroundings (Shah et al. 2021; Ossola et al. 2021; Sun et al. 2020; Sikorska et al. 2020). In addition, we have distinguished built-up areas in these buffers based on the UA2018 database. These included buildings, roads and infrastructure. These areas are poor in biologically active surface and are largely impermeable to rainwater. We did so with the knowledge that each database gives slightly different results – such situation is highlighted by Feltynowski, Kronenberg, Bergier, Kabisch, Łaszkiewicz and Strohbach (2018). To obtain results for the entire surroundings of each site, including those close to or on the border of Lublin, we added a 300-metre buffer around the administrative border of the city to the study area.

2.3.1. Expert assessment

After a study of publications and authors team discussions, we developed criteria for the expert assessment of the site’s landscapes. We did not determine a minimum size, the smallest area under investigation was 0,23 ha and the largest 86,14 ha. The expert assessment was carried out in the field from 2016 to 2018 by members of authors team in different seasons and weather conditions. Some indicators were quantitative, on a 5-1 scale, similar to the methods used by Carver et al. (2012) and Müller et al. (2015). We adopted the following categories for expert assessment:

landscape contrast, understood as a visual contrast with the surrounding area (Ode et al. 2008; Kerebel et al. 2019; Rink and Emmrich 2005). This feature can increase the attractiveness of an enchanted natural place – for example, when we are surprised by spontaneous greenery in close proximity to buildings: 5 – very clear difference, contrast with anthropogenic elements: e.g. buildings, roads, very clear differences in landform; 4 – clear difference with surrounding natural elements e.g. open area/wooded area, clear contrast in landform; 3 – noticeable differences in land cover – for example, height of vegetation, noticeable differences in landform; 2 – minor differences in nature of cover e.g. different forest type, no terrain differences; 1 – no apparent difference between the place and its surroundings;

naturalness, understood as the similarity of the growing vegetation to the potential vegetation (Carver et al. 2012; Ode et al. 2008; Ode et al. 2009; Tveit et al. 2006; Kowarik 2018) found in Lublin (Matuszkiewicz 2008): 5 – very similar vegetation to natural communities, multi-layered forest communities; 4 – tall vegetation, trees, self-sown plants, many of which are species typical for the habitat; 3 – mixed areas, shrubland, low trees, bushes, grassland vegetation, meadows; 2 – area covered mostly with grassland vegetation, extensively used, rarely mown; 1 – intensively used area, mown;

uniqueness, understood as the frequency of the occurrence of a given landscape type in the studied area (Ode et al. 2008; Tveit et al. 2006; Pietrzyk-Kaszyńska et al. 2017); the indicator was verified after experts assessed all the sites: 5 – unique sites (found once in the city); 4 – very rare sites (found 2-3 times); 3 – rare sites (found 4-6 times); 2 – common sites (7-10 times); 1 – very common sites (occurring more than 10 times).

In addition, we added an indicator expressed quantitatively: usage: (5) none, (4) single path, (3) single road, (2) several paths, (1) dense path network and an indicators expressed descriptively: main type of landform: flat, slope, gorge, valley, valley bottom, escarpment, plateau, hill, type of landscape composition: enclosure, massif, exposed, unclear, and for enclosure also the degree of view obstruction – expressed in % of the enclosure (Ashihara 1970; Bogdanowski, Łuczyńska-Bruzda and Novák 1979; Tveit et al. 2006; Ode et al. 2008) and main type of landcover: trees, shrubs, grass-herbaceous, mixed. We also added features that are not assigned a scale but only +/- values, following the guidance of Kerebel et al. (2019) – instead of talking about beauty in general, it is easier to break it down into understandable elements: low number of human artefacts in sight (Müller et al. 2018); interesting landform (Carver et al. 2012; Kerebel et al. 2019; Ode et al. 2008); presence of water (Ode et al. 2008); diversity of plant forms (Kerebel et al. 2019; Ode et al. 2008; Tveit et al. 2006); diversity of plant height (Kerebel et al. 2019; Ode et al. 2009); presence of old trees (Kerebel et al. 2019); screened-off, hidden place (Rink and Emmrich 2005); “visual access” to the area from the outside (Kowarik 2018) and distant view from the site, added after authors team discussion. We also added the sites’ names, which point to the type of the area and define its uniqueness – for example, shrubs, thicket, oxbow lake, etc. On the basis of the completed table (Tab.1), the expert also provided a subjective final evaluation general expert assessment, on a scale of 1-5. The expert awarded a score based on the average of the naturalness, landscape contrast, uniqueness scores; however, it could be lower or higher by 2. When indicating the overall assessment, the expert took into account other indicators. If a site received an general assessment of 5, 4 and 3, it was referred to as an ENP – an enchanted natural place. For the purposes of the study, we formulated a definition of ENP as a place within an urban area, where the landscape takes on particularly attractive forms approximating naturalness, with minimum human interference, where we can experience a sense of naturalness and which is clearly distinguishable from the urban surroundings.

We adopted the following rating scale of the general expert assessment for the study sites:

5 – an outstanding site, very attractive for a number of reasons and characteristics, with a high degree of naturalness – enchanted natural place (ENP);

4 – very attractive, very rare, highly natural area with several important features - ENP;

3 – valuable area, occurs rarely or often in the city, with high or medium naturalness, typical ENP for the city;

2 – spontaneous green area, IGS, has some features of an enchanted natural place but does not qualify as a full ENP.

1 – semi-natural green area, IGS, has no features of an ENP.

The contents of the evaluation sheet are shown in Tab.1.

Tab.1 Contents of the evaluation sheet

No.

ENP surface

surface of the buffer

built-up land in the buffer

naturalness

landscape contrast

uniqueness

main landform type

main type of landcover

landscape composition

% of enclosure

low number of human artefacts in sight

interesting landform

presence of water

distant view

diversity of plant forms

diversity of plans height

presence of old trees

screened-off, hidden place

"visual access" from the outside

usage

name

general expert assessment

m²

1-5

F-flat

S-slope

G-gorge

V-valley

B-valley bottom

E-escarpment

P-plateau

H-hill

T-trees

S-shrubs

H-grass/ herbaceous

M-mixed

E -enclosure

M - massif

X - exposed

U - unclear

+/-

5 -none

4 - single path

3 - single road

2 - several

paths

1 - dense path

network

1-5

2.3.2. Spectral indexes

Based on the Landsat image channels, we calculated spectral indexes to determine the status and biological activity of the studied area. We adopted the following spectral indexes: LST – Land Surface Temperature – to assess temperature differences (Cheng et al. 2019; Shah et al. 2021; Ossola et al. 2021); NDVI – Normalized Difference Vegetation Index – to assess biomass quantity, plant health and vigour (Rouse, Haas, Schell and Deering 1973; Yang et al. 2017; Shah et al. 2021; Kimm and Ryu 2015; Wang et al. 2005; Sikorska et al. 2020); NDMI – Normalized Difference Moisture Index – to assess the level of water stress in plants (Wilson and Sader 2002; Zheng et al. 2019; Zhu et al. 2019; Wang et al. 2005); LAI – Leaf Area Index – to determine the extent to which plants make use of light (Hauser et al. 2021; Sun et al. 2020; Kimm and Ryu 2015; Opik, Rolfe and Willis 2005).

In the next step, we calculated the mean values of each spectral indexes (NDVI, NDMI, LAI, LST) and the slope inside the polygons representing the boundaries of the ENPs and the polygons forming buffers, without ENPs. Then we compared the resulting indexes with the general expert assessment and the scores for landscape contrast, naturalness, and uniqueness.

We surveyed 91 sites using expert methods. These sites cover a total area of 634,9 ha, which is 4,3% of the city area. Each site received a general and detailed expert assessment. The number of sites in each general assessment field is fairly even. A total of 61 sites received ratings of 5, 4 and 3 – we identified these as ENPs – enchanted natural places. We identified the remaining 30 sites as “not ENP”. Only three sites received the highest naturalness rating. These sites also received the lowest landscape contrast scores. These are the most natural forest fragments, surrounded by woodland, with an arable character. No site received the lowest score of 1 out of naturalness. In terms of total area, the highest general assessment sites occupy the most land. These groups also include the largest of the areas. The number of sites that received each assessment is shown in Table 2. We noted dependencies between expert assessment, spectral indexes and features shown by digital data. The relationships are presented in the Table 3 and Figs. 4–9. The locations of surveyed sites are presented on Fig.1. The fragment of the city map with surveyed areas with 300-metre buffers and built-up land in the buffer is presented on Fig.2. Examples of most visually attractive sites identified as ENPs are presented on Fig.3. Table with full data (Appendix A) and shapefiles (Appendix B) are added in supplementary materials.

Tab.2 Quantitative summary of the sites that have received individual expert assessments
number of sites:
	Grade	general assessment (surface)	naturalness	landscape contrast	uniqueness
ENP	5	23 (278,3 ha)	3	22	17
ENP	4	21 (236 ha)	25	38	18
ENP	3	17 (58,4 ha)	44	21	20
not ENP	2	21 (43,8 ha)	19	8	6
not ENP	1	9 (18,4 ha)	0	2	30
All		91 (634,9 ha)

3.1. Percentage of built-up area in the buffer

We found that the sites with a higher general assessment have less built-up land in the buffer and the areas with the highest scores for naturalness have a very small share of developed land in the buffer; for the others, the average share of development was found to be similar. Moreover a clear positive correspondence exists between landscape contrast and the percentage of development in the buffer. In terms of uniqueness, ENPs classified as very rare (4) and the most common (1) have more development, while unique (5) and common (2) have the least.

3.2. Relationships between the expert assessments and spectral indexes

We have identified the most scenically interesting unmanaged informal green spaces areas within the city structure. Next, we performed a collation and analysis of the relationships between the expert assessments of these sites (general assessment, naturalness, landscape contrast, uniqueness) with the spectral indexes and GIS data for these sites and their buffers (LST, NDVI, LAI, slope, share of built-up land) (Tab.3, Figs. 4–9).

Tab.3 Relationships between expert assessments, spectral indexes and GIS database
	GENERAL EXPERT ASSESSMENT	NATURALNESS	LANDSCAPE CONTRAST	UNIQUENESS
% share of built-up land in the buffer zone	the higher the rating, the less development they have	only the highest rated sites stand out – they have the least development	the higher the landscape contrast, the more built-up land in the buffer	the most prominent areas have relatively little development, no clear dependencies
LST	cooler sites are rated higher but the coolest are 4; the higher rated are also slightly more temperature contrasted with the buffer	slight tendency: sites with higher scores are cooler; wildest (5) – clearly greater temperature difference with the buffer, others slight reverse trend: the lower the naturalness, the greater the LST difference with the buffer	reverse tendency – the higher the landscape contrast, the significantly warmer the site and its difference with the buffer smaller	evidently: the higher rated the area is for its uniqueness, the cooler it is
NDVI	sites which scored 4, less spectacular, have the highest NDVI; higher rated ones contrast less with the buffer, provide a sense of space	sites with the highest scores for naturalness sites (5) have the highest NDVI and significantly lower NDVI contrasts with the buffer; other naturalness groups at a similar level and the NDVI difference increases slightly with decreasing naturalness	the lowest rated sites have the highest NDVI; evidently: the smaller the landscape contrast, the smaller the difference in indexes with the buffer – the environment is more similar	highest NDVI for sites of average uniqueness (3 and 2); NDVI differences between the site and the buffer increase with decreasing uniqueness, significant difference also for group 4
LAI	sites which scored 4, less spectacular, have the highest LAI; higher rated ones contrast less with the buffer, provide a sense of space	naturalness 3 – have the highest LAI, and naturalness 5 have the lowest (but slightly) LAI, and significantly lower LAI contrasts with the buffer, the greatest difference for the areas which scored 2 for naturalness	areas with the lowest value for contrast have the lowest LAI; in the rest, those rated 4 and 3 stand out – average contrast; the smaller the landscape contrast, the smaller the LAI difference with the buffer	average sites (uniqueness 3 and 2) have the highest LAI; for decreasing scores for uniqueness the LAI difference between site and buffer increases, significant difference also for group 4
NDMI (moisture)	the most humid are areas rated 2; the higher rated ones contrast less with the buffer	sites rated higher for naturalness have slightly higher NDMI; sites with the highest score for naturalness (5) have significantly lower NDMI contrasts; for others, the humidity difference decreases as naturalness decreases	sites with the lowest landscape contrast have the highest NDMI, also 5 and 4 stand out; the smaller the landscape contrast, the smaller the NDMI difference with the buffer	sites rated 1 for uniqueness 1 – the most humid; decrease in uniqueness gives an increase in the difference between the site and the buffer, significant difference also for group 4
Slope	higher rated sites have a greater slope and index difference between the terrain and the buffer, rating slightly 2 stands out	the highest naturalness sites (5) have the lowest slope and the smallest difference with the buffer; the others inversely – the lower the naturalness the lower the slope and the smaller the difference; the highest value and slope difference for the sites rated 4	higher landscape contrast means higher slope and greater difference between the site and the buffer; the lowest slope and the smallest index difference are for 2	sites with uniqueness 5, 4 and 3 have the highest slope and the greatest difference with the buffer, clearly distinguishable from the sites rated 2 and 1

The results show that sites rated higher in the general assessment are significantly cooler: the lowest average temperature is in areas rated 4, indicating their value for site cooling (Fig.5, Tab.3). The difference in mean temperature between sites 4 and 1 is about 1.79 °C. What is more in terms of differences between the site and its buffer, it can be seen that the highest-rated sites in the general assessment column are slightly more temperature contrasted with the buffer, these differences ranging from 1.2 to 1.0 °C. There is a noticeable but not very strong tendency between naturalness and temperature – areas rated higher are cooler and sites with the highest scores for naturalness (5) have a significantly larger temperature difference with the buffer. For the other sites, there is a slight reverse trend – the lower the score for naturalness, the slightly larger the temperature difference compared with the buffer. Moreover for the landscape contrast there is a clear trend – the larger the contrast, the warmer the site, and the slightly less distinct it is from its surroundings – the contrast is visually attractive but contributes to the warming of the ENP and increases the impact of the surroundings. Sites with higher contrast ratings have a smaller temperature difference compared with the surrounding buffer. Thus, the visual contrast – intriguing for visitors – may contribute to the heating of the ENPs. At the same time, the large temperature difference for the least visually contrasted areas is noteworthy. Analysis of these cases indicates that they are of low temperature, and their surroundings include, among others, warming farmland. The relationship between temperature and uniqueness is apparent – the most unique (such as shaded gullies, river meanders, etc.), are on average 2.62 °C cooler than the most common ENPs (usually overgrown with shrubs). In terms of differences with the temperature buffer, the biggest differences are for ratings 5 (they are also the coolest), 3 and 2 (because their surroundings are very warm).

We found that the sites rated 4 in the general assessment section have the highest NDVIs (Fig.6, Tab.3) – they are less prominent, there are no exposed slopes, water etc., there are more shrubs which obscure the view and reduce the view of the site – but they are valuable. Moreover those rated higher in the general assessment section contrast less with the buffer. The sites with the highest score for naturalness (5) have a slightly highest NDVI and significantly lower NDVI contrasts with the buffer, while NDVI is at a similar level and the NDVI difference increases slightly for other groups in the naturalness section. The lowest score for landscape contrast almost always translates into the highest one for naturalness, these areas also have the highest NDVI, while among other groups the sites rated as 4 and 3 stand out – they are averagely visually contrasted with their surroundings. The smaller the score for landscape contrast, the smaller the difference in NDVI between the site and the buffer – the surroundings and the area are more similar. The highest NDVI values characterise areas rates 3 and 2 for uniqueness and the difference in NDVI between the site and the buffer increases as the uniqueness decreases; a significant difference is also found for group 4 – very rare but not outstanding ENPs.

The results show that the highest LAIs distinguish sites rated 4 in the general assessment section, thus indicating their value and those rated higher in the general assessment section contrast less with the buffer (Fig.7, Tab.3). What is more the highest LAI was recorded for areas rated 3 for their naturalness – intermediate, mostly covered with a mixture of trees, shrubs, open areas; this index measures foliage and indicates that these areas have value in terms of vegetation density and amount of biomass. The lowest, albeit slightly, LAI occurred for the highest naturalness. The sites rated the highest for naturalness (5) have smaller LAI contrasts with the buffer; this difference is markedly larger for remaining areas, and the highest for those rated with 3. The areas with the lowest landscape contrast – almost identical to the highest naturalness, have the lowest LAI; while among the other groups the areas rated as 4 and 3 stand out – with an average visual contrast with their surroundings. Moreover there is a clear relationship between the difference in LAI indexes and the buffer in the case of landscape contrast – the lower the contrast, the smaller the difference in indexes – the environment is more similar. The highest LAI values have areas of average uniqueness 3 and 2 and differences between terrain and buffer in LAI increase as uniqueness decreases. Group 4, very rare but not outstanding ENP, also stand out.

In terms of NDMI – the sites with a general assessment of 2 stand out clearly thus indicating their value for water storage (Fig.8, Tab.3). Those rated higher in the general assessment contrast less with the buffer. Moreover rated higher in naturalness – have slightly higher NDMIs. Sites rated as most for naturalness (5) have significantly lower NDMI contrasts with the buffer; for subsequent groups the humidity difference decreases. Sites with the lowest landscape contrast have the highest humidity ratings. Areas rated 5 and 4 also stand out. What is more there is a clear relationship between the difference in NDMI indexes with the buffer and landscape contrast – the lower the contrast, the smaller the difference in indexes – the environment is more similar. The most common sites in terms of uniqueness (1) are also the most humid, indicating that what is average and frequent in the city is valuable in terms of humidity, even if it does not stand out in terms of temperature. Furthermore differences between the terrain and the buffer in humidity increase as uniqueness decreases; group 4 also stands out here.

There is a clear trend for the general assessment – the higher the assessment, the greater the slope and the difference in rate between the terrain and the buffer (Fig.9, Tab.3). Only the sites rated 2 slightly stand out – the analysis of these cases shows that the condition of the landform was not sufficient here for attractiveness and the impression of naturalness. However the areas with the highest naturalness (5) are the least varied in terms of slope; they are flat or slightly undulating, but the trend is clear for the subsequent groups – the higher the naturalness value, the more varied the terrain. Attractive escarpments, ravines and valley slopes dominate in this group (4). The distribution of the slope difference between the terrain and the buffer is similar; those rated as most natural contrast the least, while those rated 4 again stand out clearly. Moreover the relationship between slope and landscape contrast is also evident – the areas with the highest contrast are also the most sculpted. The relationship with the slope difference between the terrain and the buffer is similar – the least sculpted and at the same time the least differentiated from the surroundings are the areas rated 2 in the landscape contrast. The areas rated 5, 4 and 3 in uniqueness have the most varied topography, clearly distinguishing them from the flatter and more common sites; especially area 4 has the highest index and the greatest difference of this index with the buffer; This is due to the fact that this group includes areas that are rare but nevertheless occur 2 or 3 times in the city – for example, inaccessible slopes or ravines.

With regard to the method of site designation, we used an expert method based on indicators selected from the literature. In assessing the sites, we elected to use simple indicators that will be replicable in other cities, tailored to their specific characteristics. Of the numerous indicators for the landscape as a whole, we selected those that were relevant to the concrete situation addressed (Ode et al. 2008), focused on positive factors, and did not introduce aspects of ugliness (“disturbances”) that Kerebel et al. (2019) used.

The result is a simple, reproducible tool that can be helpful when working with local authorities, NGOs or educational institutions in other cities, as advocated by Kremer et al. (2016), although it requires considerable fieldwork. We also proposed that the areas with the highest expert assessment be referred to as enchanted natural places (ENPs) to emphasise their uniqueness in the cityscape. The distribution of expert ratings was mostly balanced (Tab.2). However, due to the nature of the city, the highest expert assessment of naturalness was a rare feature, as was the lowest degree of contrast – at the same time these were the same areas, which may distort the results for these groups. At the same time, this indicates that urban wildness is always relative, perceived in the context of the city (Müller et al. 2018). Moreover we used a single, fixed buffer of 300 m as the areas closest to the site to capture the relationship between the areas and their surroundings. Similar methods were used by Shah et al. (2021), Ossola et al. (2021) and Sun et al. (2020), among others, although we did not investigate situations where buffers overlapped with neighbouring study sites or buffers, as pointed out by Sun et al. (2020) – the result of which may be an unrecognised effect of mutual proximity of the sites on study results. We observed a number of trends in the relationships of expert assessments and spectral indexes, although we are aware of the limitations of the method adopted, the indicators selected and the use of average values for comparisons that do not reveal the nuances of each site.

In the overall assessment we noted that highly valued sites (ENPs) tended to have more favourable, less built-up surroundings. We observed clear relationships between temperature and expert assessments. Sites rated higher were generally cooler. However, the lowest temperature and highest NDVI and LAI were in areas rated as 4 – less outstanding although very valuable. These areas had lower NDVI, LAI, NDMI contrasts with their surroundings. Similar relationships between low temperature and high NDVI were observed by Yang et al. (2017), Zheng at al. (2019), Shah et al. (2021) and Chen et al. (2022).

An interesting observation was made concerning the landscape contrast, which was clearly related to the proportion of built-up areas in the buffer. Where these values were higher, there was a greater temperature and a lower temperature difference with the surroundings. The reason for this is most probably the larger impact of the built-up buffer on the site. The study by Ode et al. (2009) does not clearly indicate which type of land boundary is considered more natural. In our research we found contrast to be an attractive feature, and thus visual contrast, intriguing to users, may contribute to ENPs heating. These observations are complemented by the results of researchers such as Shah et al. (2021), Ossola et al. (2021), Sun et al. (2020), Yang et al. (2017), who highlight the role of even small areas in neighbourhood cooling.

There is a clear relationship between temperature and uniqueness – the most unique, such as shaded gorges, river meanders, etc., were much cooler than areas overgrown with shrub common in the city. The results as a whole confirm the role of land for temperature reduction – what Cheng et al. (2019), among others, refer to as UCI – “urban cooling islands”. Intriguing results were obtained for the LAI, which is sometimes shaped differently from the NDVI. The dissimilarity of the two indexes was pointed out, among others, by Wang et al. (2005). For example, the highest LAI values and contrast with the buffer have sites defined as having average naturalness (3). This result indicates that the sites have value in terms of vegetation density and amount of biomass. In contrast, the sites with the highest naturalness (5) had slightly but nevertheless the lowest LAI. This result stands in contrast to the conclusions of Hauser et al. (2021), who found generally higher LAI values for forests than for shrubland. Similar findings were also presented by Sun et al. (2020), who, by examining the Leaf Area (LA) parameter derived from the LAI, showed that the highest values of the index (and also the lowest temperature) were obtained in forested areas. These researchers also derived the LAI from in situ studies, and additionally Kimm and Ryu (2015) indicated significant seasonal variability in LAI. This presents a challenge for further research.

Clear relationships are evident between slope and expert assessments. In our study we used a simple indicator – the average slope. Other researchers adopt different, more specific indicators, but the basic conclusions are similar and confirm that varied landform supports visual attractiveness: (Müller et al. 2018; Müller et al. 2015; Carver et al. 2012). Interestingly, the highest naturalness values (forest fragments) had the lowest slope. Also the areas lowest rated in general assessment had a slightly higher index – indicating that the landform alone was not sufficient for a high rating and impression of naturalness.

In the results obtained, the expert assessments were not always confirmed directly in the spectral indexes/GIS data. This is particularly evident in the general assessment, where the highest NDVI and LAI values and the lowest temperature were in areas assessed as 4 – they are less prominent but probably have more biomass, there are no exposed slopes, water etc., there are more shrubs which obscure the view and reduce the view of the area. In terms of NDMI (moisture), the sites rated 2 stand out. This indicates their role in water storage; however, this result is puzzling as it does not correlate with the temperature of the land, which is considerably warmer. According to researchers such as Zheng et al. (2019) and Zhu et al. (2019), higher LST should correlate with lower NDMI.

As far as uniqueness is concerned, the highest moisture content is found in sites rated 1 – the most common “bushes and shrubs”. Similarly for NDVI and LAI, sites 3 and 2 – averagely unique – achieved the highest values. We referred to this phenomenon as the eulogy to mediocrity. In our perception we pay attention to the unique landscape of the enchanted natural places but it is also worth appreciating what seems ordinary. A similar situation, where less wild areas tend to be richer in species, has been pointed out by Müller et al. (2018). Also the intensity of use did not clearly result in a lower rating. This is consistent with the observations of Rupprecht and Byrne (2014), who notes that, “residents also prefer a certain level of maintenance (a ‘tended’ look, cleanliness)”. The analysed sites are largely unused and little known to the local community despite the proximity of the built-up areas – hence at this stage we decided not to carry out a survey. Experiences in this field by authors such as Rupprecht et al. (2015), Kendal et al. (2015), Sikorska et al. (2020), Włodarczyk - Marciniak et al. (2020) may be helpful for further research, including one on changes in use in the context of increased awareness of urban unmanaged IGS and conscious sharing. Rink and Emmrich (2005) pointed out that appropriate and limited access contributes to land conservation. Further research directions could be: to expand research to include nature inventories – for example, a bioscore (Müller et al. 2018), to recognise the impact of site size on expert and spectral indexes, to expand research to other comparison sites, the impact of noise (Müller et al. 2018), the impact of distance from the centre, relationships with residential areas and accessibility, ecosystem services (Kerebel et al. 2019), planning considerations, possible conservation directions and the creation of “new wilderness” (Henne 2005).

Following the statement made by Pietrzyk-Kaszyńska et al. (2017) about the need for further recognition of IGS and their inclusion in urban management policies, as well as the postulates of Kremer et al. (2016) concerning the study of human-nature relations, we conducted a study of the most visually attractive informal green spaces based around the Lublin area.

We have developed an expert assessment method to identify the most attractive areas, distinguish them from other unmanaged IGS and proposing the name “enchanted natural places” (ENP) for these sites. With the knowledge that each city has its own specifics (Kremer et al. 2016), this method can be adapted to study other cities. We juxtaposed the results of the expert assessment with spectral indexes (LST, NDVI, LAI, NDMI) and GIS data and found a number of relationships between them. Analyses for the areas and a 300 m buffer around them were performed. We found clear temperature relationships: the temperature of the sites decreases as the overall attractiveness increases, the temperature of the sites increases as the perceived landscape contrast with the surroundings increases. There is also a clear relationship between landscape contrast and the difference in area/buffer ratios. We also found a relationship between the expert assessments and the land slope. The spectral indexes largely confirm the results of the expert assessment, but it should be noted that NDVI, LAI and NDMI also indicated the values of the sites assessed lower by the experts. Hence, we conclude that land conservation should value not only the scenically spectacular informal green spaces but also average “bushes and shrubs”. Both approaches used reveal important features of the sites hence it is beneficial to combine direct ground survey methods and spectral analyses. Identifying the most attractive areas – ENPs, “giving an identity”, can be the first step to formally protect the diverse urban wilderness and the presented method based on simple perceived indicators and spectral indexes can be helpful in the dialogue with local authorities and residents. Their protection may start with the most visually interesting fragments, but it should gradually extend to more common, but also valuable, areas.

DECLARATION OF COMPETING INTEREST

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

ACKNOWLEDGEMENTS

We would like to express our gratitude to Ms Aleksandra Piasecka who translated our article into English and Steve Jones who made proofreading.

Ashihara Y (1970) Exterior Design in Architecture. Van Nostrand Reinhold Company, New York
Bogdanowski J, Łuczyńska-Bruzda M, Novák Z (1979) Architektura krajobrazu (Landscape architecture). PWN, Warszawa – Kraków
Carver S, Comber A, McMorran R, Nutter S (2012) A GIS model for mapping spatial patterns and distribution of wild land in Scotland. Landscape and Urban Planning 104:395-409. https://doi.org/10.1016/j.landurbplan.2011.11.016
Chen H, Deng Q, Zhou Z, Ren Z, Shan X (2022) Influence of land cover change on spatio-temporal distribution of urban heat island — a case in Wuhan main urban area. Sustainable Cities and Society 79. https://doi.org/10.1016/j.scs.2022.103715
Cvejić R, Eler K, Pintar M, Železnikar Š, Haase D, Kabisch N, Strohbach MW (2015) A typology of urban green spaces, ecosystem services provisioning and demands. https://www.e-pages.dk/ku/1334/html5/. Accesed 1 July 2022
Cheng L, Guan D, Zhou L, Zhao Z, Zhou J (2019) Urban cooling island effect of main river on a landscape scale in Chongqing, China. Sustainable Cities and Society 47. https://doi.org/10.1016/j.scs.2019.101501
Feltynowski M, Kronenberg J, Bergier T, Kabisch N, Łaszkiewicz E, Strohbach M (2018) Challenges of urban green space management in the face of using inadequate data. Urban Forestry & Urban Greening 31:56-66. https://doi.org/10.1016/j.ufug.2017.12.003
Florgård C (2007) Preserved and remnant natural vegetation in cities: A geographically divided field of research. Landscape Research 32(1):79-94. https://doi.org/10.1080/01426390601097750
Hauser LT, Féret J-B, An Binh N, van der Windt N, Sil ÂF, Timmermans J, Soudzilovskaia NA, van Bodegom PM (2021) Towards scalable estimation of plant functional diversity from Sentinel-2: In-situ validation in a heterogeneous (semi-)natural landscape. Remote Sensing of Environment 262. https://doi.org/10.1016/j.rse.2021.112505
Henne SK (2005) “New Wilderness” as an Element of the Peri-Urban Landscape. In: Kowarik I, Körner S (eds) Wild Urban Woodlands. Springer-Verlag, Berlin – Heidelberg, pp 247-262. https://doi.org/10.1007/3-540-26859-6_15
Kendal D, Ford RM, Anderson NM, Farrar A (2015) The VALS: A new tool to measure people's general valued attributes of landscapes. Journal of Environmental Management 163:224-233. https://doi.org/10.1016/j.jenvman.2015.08.017
Kerebel A, Gélinas N, Déry S, Voigt B, Munson A (2019) Landscape aesthetic modelling using Bayesian networks: Conceptual framework and participatory indicator weighting. Landscape and Urban Planning 185:258-271. https://doi.org/10.1016/j.landurbplan.2019.02.001
Kowarik I (2005) Wild Urban Woodlands: Towards a Conceptual Framework. In: Kowarik I, Körner S (eds) Wild Urban Woodlands. Springer-Verlag, Berlin – Heidelberg, pp 1-32. https://doi.org/10.1007/3-540-26859-6_1
Kowarik I (2018) Urban wilderness: Supply, demand, and access. Urban Forestry & Urban Greening 29:336-347. https://doi.org/10.1016/j.ufug.2017.05.017
Kimm H, Ryu Y (2015) Seasonal variations in photosynthetic parameters and leaf area index in an urban park. Urban Forestry & Urban Greening 14(4):1059-1067. https://doi.org/10.1016/j.ufug.2015.10.003
Kremer P, Hamstead Z, Haase D, McPhearson T, Frantzeskaki N, Andersson E, Kabisch N, Larondelle N, Lorance Rall E, Voigt A, Baró F, Bertram C, Gómez-Baggethun E, Hansen R, Kaczorowska A, Kain J-H, Kronenberg J, Langemeyer J, Pauleit S, Rehdanz K, Schewenius M, van Ham C, Wurster D, Elmqvist T (2016) Key insights for the future of urban ecosystem services research. Ecology and Society 21(2):29. http://dx.doi.org/10.5751/ES-08445-210229
Matuszkiewicz JM (2008) Potencjalna roślinność naturalna Polski (Potential natural vegetation of Poland). IGiPZ PAN, Warszawa
Müller A, Bøcher PK, Fischer C, Svenning J-C (2018) ‘Wild’ in the city context: Do relative wild areas offer opportunities for urban biodiversity?. Landscape and Urban Planning 170:256-265. https://doi.org/10.1016/j.landurbplan.2017.09.027
Müller A, Bøcher PK, Svenning J-C (2015) Where are the wilder parts of anthropogenic landscapes? A mapping case study for Denmark. Landscape and Urban Planning 144:90-102. https://doi.org/10.1016/j.landurbplan.2015.08.016
Ode Å, Tveit MS, Fry G (2008) Capturing Landscape Visual Character Using Indicators: Touching Base with Landscape Aesthetic Theory. Landscape Research 33(1):89-117. https://doi.org/10.1080/01426390701773854
Ode Å, Fry G, Tveit MS, Messager P, Miller D (2009) Indicators of perceived naturalness as drivers of landscape preference. Journal of Environmental Management. 90(1):375-383. https://doi.org/10.1016/j.jenvman.2007.10.013
Opik H, Rolfe SA, Willis AJ (2005) The Physiology of Flowering Plants, Cambridge University Press, New York
Ossola A, Jenerette GD, McGrath A, Chow W, Hughes L, Leishman MR (2021) Small vegetated patches greatly reduce urban surface temperature during a summer heatwave in Adelaide, Australia. Landscape and Urban Planning 209. https://doi.org/10.1016/j.landurbplan.2021.104046
Pietrzyk-Kaszyńska A, Czepkiewicz M, Kronenberg J (2017) Eliciting non-monetary values of formal and informal urban green spaces using public participation GIS. Landscape and Urban Planning 160:85-95. https://doi.org/10.1016/j.landurbplan.2016.12.012
Rink D, Emmrich R (2005) Surrogate Nature or Wilderness? Social Perceptions and Notions of Nature in an Urban Context. In: Kowarik I, Körner S (eds) Wild Urban Woodlands. Springer-Verlag, Berlin – Heidelberg, pp. 67-80. https://doi.org/10.1007/3-540-26859-6_4
Rouse JW, Haas RH, Schell JA, Deering DW (1973) Monitoring the vernal advancement and retrogradation (green wave effect) of natural vegetation. Remote Sensing Center, Texas A&M University
Rupprecht C, Byrne JA (2014) Informal urban greenspace: A typology and trilingual systematic review of its role for urban residents and trends in the literature. Urban Forestry & Urban Greening 13(4):597-611. https://doi.org/10.1016/j.ufug.2014.09.002
Rupprecht C, Byrne JA, Ueda H, Lo AY (2015) ‘It’s real, not fake like a park’: Residents’ perception and use of informal urban green-space in Brisbane, Australia and Sapporo, Japan. Landscape and Urban Planning 143:205-218. https://doi.org/10.1016/j.landurbplan.2015.07.003
Sikorska D, Łaszkiewicz E, Krauze K, Sikorski P (2020) The role of informal green spaces in reducing inequalities in urban green space availability to children and seniors. Environmental Science & Policy 108:144-154. https://doi.org/10.1016/j.envsci.2020.03.007
Shah A, Garg A, Mishra V (2021) Quantifying the local cooling effects of urban green spaces: Evidence from Bengaluru, India. Landscape and Urban Planning 209. https://doi.org/10.1016/j.landurbplan.2021.104043
Sun X, Tan X, Chen K, Song S, Zhu X, Hou D (2020) Quantifying landscape-metrics impacts on urban green-spaces and water-bodies cooling effect: The study of Nanjing, China. Urban Forestry & Urban Greening 55. https://doi.org/10.1016/j.ufug.2020.126838
Taylor L, Hochuli DF (2017) Defining greenspace: Multiple uses across multiple disciplines. Landscape and Urban Planning 158:25-38. https://doi.org/10.1016/j.landurbplan.2016.09.024
Tveit M, Ode Å, Fry G (2006) Key concepts in a framework for analysing visual landscape character. Landscape Research 31(3):229-255. https://doi.org/10.1080/01426390600783269
Wang Q, Adiku S, Tenhunen J, Granier A (2005) On the relationship of NDVI with leaf area index in a deciduous forest site. Remote Sensing of Environment 94(2):244-255. https://doi.org/10.1016/j.rse.2004.10.006
Wilson EH, Sader SA (2002) Detection of forest harvest type using multiple dates of Landsat TM imagery. Remote Sensing of Environment 8(3):385-396. https://doi.org/10.1016/S0034-4257(01)00318-2
Włodarczyk-Marciniak R, Sikorska D, Krauze K (2020) Residents’ awareness of the role of informal green spaces in a post-industrial city, with a focus on regulating services and urban adaptation potential. Sustainable Cities and Society 59. https://doi.org/10.1016/j.scs.2020.102236
Yang C, He X, Wang R, Yan F, Yu L, Bu K, Yang J, Chang L, Zhang S (2017) The Effect of Urban Green Spaces on the Urban Thermal Environment and Its Seasonal Variations. Forests 8(5):153. https://doi.org/10.3390/f8050153
Zevenbergen LW, Thorne CR (1987) Quantitative analysis of land surface topography. Earth Surface Processes and Landforms 12:47-56. https://doi.org/10.1002/esp.3290120107
Zheng H, Chen Y, Pan W, Cai Y, Chen Z (2019) Impact of Land Use/Land Cover Changes on the Thermal Environment in Urbanization: A Case Study of the Natural Wetlands Distribution Area in Minjiang River Estuary, China. Polish Journal of Environmental Studies 28(4):3025-3041. https://doi.org/10.15244/pjoes/93743
Zhu X, Wang X, Yan D, Liu Z, Zhou Y (2019) Analysis of remotely-sensed ecological indexes' influence on urban thermal environment dynamic using an integrated ecological index: a case study of Xi’an, China. International Journal of Remote Sensing 40(9):3421-3447. https://doi.org/10.1080/01431161.2018.1547448
USGS https://www.usgs.gov/core-science-systems/nli/landsat/landsat-8?qt-science_support_page_related_con=0#qt-science_support_page_related_con. Accesed 1 July 2021 Remote sensing data and digital data:
EarthExplorer https://earthexplorer.usgs.gov/ Accesed 1 July 2022
Catalyst.Earth https://catalyst.earth/products/catalyst-pro/ Accesed 1 July 2022
Copernicus https://land.copernicus.eu/local/urban-atlas Accesed 1 July 2022
National Geoportal of Poland www.geoportal.gov.pl Accesed 30 August 2018
Statistics Poland, Local Data Bank https://bdl.stat.gov.pl Accesed 30 August 2018

No competing interests reported.

Download PDF

Version 2

posted

You are reading this latest preprint version

Using expert and spectral methods to assess visually attractive urban informal green spaces in Lublin, Poland.

Status:

Version 2

Abstract

Figures

Introduction

Methods

Results

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 2