Urban blue–green space landscape ecological health assessment based on the integration of pattern, process, function and sustainability
Study areaHarbin is located in the centre of Northeast Asia, between 44°04’–46° 40′ N and 125° 42′–130° 10′ E24,26. The site has a mid-temperate continental monsoon climate, with an average annual temperature of 3.6° C and an average annual precipitation is 569.1 mm. The main precipitation months being from June to September, accounting for about 60% of the annual precipitation, the main snow months are from November to January24,25. The overall topography is high in the east and low in the west, with mountains and hills predominating in the east and plains predominating in the west27. In this study, we identified the central district of Harbin, where urban construction activities are frequent and the population is dense, as the study area. According to the “Harbin City Urban Master Plan (2011–2020)” (revised draft in 2017), the specific scope includes Daoli District, Daowai District, Nangang District, Xiangfang District, Pingfang District, Songbei District’s administrative district, Hulan District, and Acheng District part of the area, with a total area of 4187 km2 (Fig. 2). The blue–green space in this study included woodland, grassland, cultivated land, wetland and water that permeate inside and outside the construction sites. They all have integrated functions such as ecology, supply, beautification, culture, and disaster prevention and avoidance, and have a decisive influence on the urban ecological environment.Figure 2Schematic of study area. The Figure is created using ArcGIS ver.10.2 (https://www.esri.com/).Full size imageData sourcesThe data used in this research included the following: land-cover date (30 m × 30 m) of two periods (2011, 2020) spported by the China Geographic National Conditions Data Cloud Platform (http://www.dsac.cn/), Meteorological datasets (1 km × 1 km) were obtained from the Resource and Environmental Science Data Center of the Chinese Academy of Sciences (http:∥www.resdc.cn/), including air temperature, precipitation, and surface runoff. ASTER GDFM elevation data (30 m × 30 m) came from the Geospatial Data Cloud (http:∥www.gscloud.cn), from which the slope was extracted. Soil data (1 km × 1 km) were from the World Soil Database (HWSD) China Soil Data Set (v1.1). The normalized difference vegetation index (NDVI) and modified normalized difference water index (MNDWI) data (30 m × 30 m) came from the National Comprehensive Earth Observation Data Sharing Platform (http://www.chinageoss.org/), ET datasets (30 m × 30 m) were drawn from the NASA-USGS (https://lpdaac.usgs.gov/). Social and economic data were mainly obtained through the Harbin statistical yearbook and the Harbin social and economic bulletin.Framework of urban blue–green space LEH assessmentUrban blue–green space is a politically defined man-land coupling region composed of ecological, economic, and social systems, which is greatly disturbed by human activities11. The essence of urban blue–green space LEH is that the landscape ecological function sustainably meets human needs28,29. The landscape ecological function reflects the value orientation of human beings to blue–green space, and to a large extent affects the blue–green landscape ecological pattern and process. The interaction between the blue–green landscape ecological pattern and process drives the overall dynamics of blue–green space. Meanwhile, presenting certain landscape ecological function characteristics, which provide ecological support for various human activities30,31,32. While the pattern and process of blue–green space both profoundly influence and are influenced by human activities33,34. This influence is long-term, the standard of LEH should not be fixed in real-time health, but should fully consider the sustainability of the health state.In summary, the landscape ecological pattern, process, function, and sustainability are not separate, but a complex of mutual integration, and organic unity. In this study, we constructed an integrated assessment framework of blue–green space LEH that included four units: pattern, process, service, and sustainability (Fig. 3). In the assessment framework, the LEH of urban blue–green space involves two dimensions: the first is the health status of the urban blue–green space itself, emphasizing the maintenance of the ecological conditions, thereby potentially satisfying a series of diversity goals. The other is that urban blue–green space, as a part of social and economic development, could sustainably provide the ability to meet (subject) needs and goals.Figure 3Key units, interactions of urban blue–green space LEH.Full size imageLandscape ecological patternThe landscape ecological pattern of urban blue–green space is a spatial mosaic combination of landscape elements at different levels or the same level. Affected by human activities interference31, the landscape ecological pattern shows the changing trend of landscape structure complexity, landscape type diversification, and landscape fragmentation. The assessment of urban landscape ecological pattern should be a comprehensive reflection of this changing trend1. Landscape pattern indexes are the most frequently applied which could reflect the structural composition and spatial configuration characteristics of the landscape4,35. This study took landscape ecology as the entry point and selected the landscape pattern indexes that can quantitatively reflect the change characteristics of landscape structural composition and spatial configuration under the disturbance. In this way, the landscape disturbance index (U), landscape connectivity index (CON), and landscape adaptability index (LAI) were used as the indexes for the assessment of landscape ecological pattern health.
(1)
Landscape disturbance index (U)
There are two kinds of relationships between the landscape ecological pattern and the external disturbance: compatibility and conflict. As the landscape ecological pattern has accommodating characteristics, the disturbance beyond the accommodating capacity will degrade the landscape ecological pattern36,37. The landscape disturbance index (U) could characterize the degree of fragmentation, dispersion, and morphological changes in landscape pattern38. The index is a comprehensive index that can reflect the health of the landscape pattern by quantifying the ability of ecosystems to accommodate external disturbances. It consists of the landscape fragmentation index, the inverse of the fractional dimension, and the dominance index. They measure the response of the landscape pattern to external disturbance from the perspective of different landscape types, the same landscape type, and landscape diversity, respectively36,38, and their weights were determined by the entropy weight method. The formula is as follows:$$ U = alpha N_{{{Fi}}} + bD_{{{Fi}}} + cD_{{{Oi}}} $$
(1)
where NFi is the landscape fragmentation index, DFi is the inverse of the fractional dimension, DOi is the dominance index, and a, b, and c are the corresponding weights, which were 0.20, 0.5, and 0.3 in this study, respectively.
(2)
Landscape connectivity index (CON)
The most direct result of landscape ecological pattern degradation caused by external disturbance is that the flow of energy, material, and information among ecological patches is reduced or even blocked, ultimately the stability of the landscape pattern is decreased. The connectivity could characterize the ability of landscape ecological pattern to mitigate risk transmission, which is significant for the dynamic stability of landscape ecological pattern39,40. The landscape connectivity index (CON) could measure the connectivity between ecosystem components through the aggregation or dispersion trend of patches41. The better the connectivity, the stronger the stability of landscape ecological pattern. The formula is as follows:$$ CON = frac{{100sumlimits_{s = 1}^{q} {sumlimits_{h ne l}^{p} {C_{{{shl}}} } } }}{{sumlimits_{s = 1}^{s} {left[ {q_{{s}} (q_{{s}} – 1)/2} right]} }} $$
(2)
where qs is the number of plaques of patch type s, Cshl is the link between patch h and patch l in s within the delimited distance.
(3)
Landscape Restorability Index (LRI)
The ability to recover to its original structure when subjected to disturbances is an important criterion for the landscape ecological pattern42. Research confirmed that the restorability of the landscape ecological pattern is closely related to the structure, function, diversity, and uniformity of distribution. The landscape restorability index (LRI) combines the above landscape information and could indicate the restorability of the landscape ecological pattern in response to disturbance43. The index consists of the patch density, Shannon diversity index, and the landscape evenness, the patch density is the number of patches per square kilometer. The Shannon diversity index reflects the change in the proportion of landscape types. The landscape evenness index shows the distribution evenness of patches in terms of area. The larger the LRI index, the more complex and evenly distributed the structure is, and the more recovery ability of the landscape pattern against disturbance is. The formula is as follows:$$ LRI = PD times SHDI times SHEI $$
(3)
where PD is the patch density, SHDI is the Shannon diversity index, and SHEI is the landscape evenness index.Landscape ecological processThe landscape ecological process of urban blue–green space is extremely complex for it involves multiple factors such as natural ecology, economy, and culture. Landscape ecological process assessment is the measure of the self-organized capacity and the efficiency of ecological processes within and among patches44. A blue–green space with a healthy landscape ecological process should have the ability to adapt to conventional land use under human management and maintain physiological integrity while maintaining the balance of ecological components. Specifically, the landscape ecological process could quickly restore its balance after severe disturbances, with strong organization, suitability, recoverability, and low sensitivity45,46. A single model hardly to gets good research on landscape ecological process under the urban scale. The comprehensive application of multidisciplinary methods is effective means to solve the problem. Regarding this, we selected ecological indexes and models from four aspects: organization, suitability, restoration, and sensitivity to assess the landscape ecological process of urban blue–green space.
(1)
Organization index (O)
The organization of the landscape ecological process is the maintenance ability of stable and orderly material cycling and energy flow within and between landscapes47. The normalized vegetation index (NDVI) and the modified normalized difference water index (MNDWI) could reflect the efficiency and order of ecological processes. Such as accumulation of organic matter, fixation of solar energy, nutrient cycling, regeneration, and metabolism13. The indexes are the external performance of the internal dynamics and organizational capabilities of the ecological process. In recent years, it has been widely used in the assessment of related to landscape ecological process. The formulas are as follows:$$ NDVI = frac{NIR – R}{{NIR + R}} $$$$ MNDWI = frac{p(green) – p(MIR)}{{p(green) + p(MIR)}} $$
(4)
where (NDVI) is the normalized vegetation index, (MNDWI) is the modified water body index, (NIR) is the reflectance value in the near-infrared band, (R) is the reflectance value in the visible channel, (p(green)) and (p(MIR)) are the normalized values in the green and mid-infrared bands.
(2)
Suitability index (Q)
The suitability of the landscape ecological process is a measurement of the self-regulating ability of the landscape ecosystem. That is, to effectively maintain the ecological process in a state of being protected from disturbance during the occasional changes caused by the external environment2. The water conservation amount index (Q) can measure the operating capacity of ecosystems to maintain ecological balance, water conservation, climate regulation, and other ecological processes by integrating the water balance of rainfall, surface runoff, and evaporation41. It could reflect the suitability of landscape ecological process to regional environment and developmental conditions. The formula is as follows:$$ Q = R – J – ET $$
(5)
where Q is the water conservation amount, R is the annual rainfall, J is the surface runoff, ET is the evapotranspiration.
(3)
Recoverability index (ECO)
The recoverability of the landscape ecological process refers to the ability of an ecosystem to return to its original operating state after being subjected to external impacts. Land-use types play an essential role in landscape ecological recoverability48. The ecological recoverability index (ECO) uses the resilience coefficients of land-use types to reflect the level of ecosystem resilience38. Based on previous studies, the resilience coefficient of land-use types was assigned (Table 1).
(4)
Sensitivity index(A)
Table 1 Resilience coefficients of different land use types.Full size tableThe sensitivity index (A) could be used to indicate landscape ecological process formation, change, and vulnerability to disturbance31. We started from the physical effects of blue–green space on sand production, water confluence, and sediment transport, introduced the Soil Erosion Modulus to characterize the sensitivity of landscape ecological processes to disturbance. The index effectively combines landscape ecology, erosion mechanics, soil science, and sediment dynamics49. The formula is as follows:$$ begin{gathered} A = R_{{i}} cdot K cdot LS cdot C cdot P hfill \ L = (l/22.1)^{m} hfill \ S = left{ begin{gathered} 10.8sin theta + 0.03,theta < 5^{ circ } hfill \ 16.8sin theta - 0.50,5^{ circ } le theta < 10^{ circ } hfill \ 21.9sin theta - 0.96,theta ge 10^{ circ } hfill \ end{gathered} right. hfill \ C = left{ begin{gathered} 1,c = 0 hfill \ 0.6508 - 0.3436lg c,0 < c le 78.3% hfill \ 0,c > 78.3% hfill \ end{gathered} right. hfill \ end{gathered} $$
(6)
where A is the soil erosion modulus. Ri is the rainfall erosion factor, K is the soil erosion factor, L and S are slope the length factor and the slope factor respectively, C is the vegetation coverage and management factor, P is the soil and water conservation factor, l is the slope length value, m is the slope length index, and θ the is slope value.Landscape ecological functionThe landscape ecological function determines the ability of ecological service50,51,52, the ecological service of urban blue–green space depends on the human value orientation48. It includes four categories: supply, support, regulation, and culture. Based on Maslow’s Hierarchy of Needs and Alderfer’s ERG theory, scholars have summarized the three major needs of human beings for urban blue–green space. Namely, securing the living environment to meet the survival needs, improving social relationships to meet the interaction needs, and cultivating cultural cultivation to meet the development needs53. Specifically corresponding to the landscape ecological function of urban blue–green space, supply is not the main function, only plays a subsidiary role, support is the basic guarantee, regulation is the basic need for urban environmental construction, and culture is an important element of high-quality social life. Ecosystem service value (ESV) can realize the measurement of ecological service function by calculating the specific value of life support products and services produced by the ecosystem54,55,56. Considering the human value orientation of the urban blue–green space landscape ecological function, the weights were given by consulting 16 experts, with supply, regulation, support, and culture weights of 0.2, 0.3, 0.3, 0.2, respectively. The formula is as follows:$$ ESV = sumlimits_{k = 1}^{n} {S_{k} times V_{k}^{{}} } $$
(7)
where Sk is the area of landscape type k, Vk is the value coefficient of the ecosystem service function of landscape type k .Landscape ecological sustainabilityWu (2013) proposed a research framework for landscape sustainability based on a summary of related studies, stating that landscape ecological sustainability is the ability to provide ecosystem services in a long-term and stable manner34. The framework emphasized that landscape sustainability should focus on the analysis of ecosystem service trade-offs effect34,57. In the process of dynamic change of urban blue–green space ecosystem, there are complex trade-offs among various ecosystem services. This is important for promoting the optimal overall benefits of various ecosystem services and achieving sustainable development of urban ecology58. In addition, as a special type of human-centered ecosystem developed by humans based on nature, human well-being is also very important for the landscape ecological sustainability of urban blue–green space. For this reason, we introduced ecosystem service trade-offs (EST) and ecological construction input (IEC) as assessment indexes of landscape ecological sustainability.
(1)
Ecosystem service trade-offs (EST)
This study applied the root mean square deviation of ecological services to quantify ecosystem service trade-offs (EST). The index could effectively measure the average difference in standard deviation between individual ecosystem services and the average ecosystem services. It is a simple and effective way to evaluate the trade-offs among ecosystem services. The formula is as follows:$$ EST = sqrt {frac{1}{n – 1}sumnolimits_{i = 1}^{n} {(ES_{std} – overline{ES}_{std} } } )^{2} $$
(8)
where ESstd is the normalized ecosystem services, n is the number of ecosystem services , and (overline{ES}_{std}) is the mean value of normalized ecosystem services.
(2)
Ecological construction input (ECI)
Human well-being is a premise for the landscape ecological sustainability of urban blue–green spaces, it is closely related to government investment in ecological construction planning34. From the perspective of economics, this study assessed the human well-being obtained by urban blue–green space with the ratio of urban ecological construction investment to GDP, that is, the ecological construction input (ECI). The formula is as follows:$$ ECI = EI/G $$
(9)
where EI is the amount of ecological construction investment, and G is the gross regional product.Evaluation methodThe index weight determines its relative importance in the index system, and the selection of the weight calculation method in the decision-making of multi-attribute problems has an important impact on the assessment results21. Traditional weighting methods can be divided into two categories, subjective weighting method and objective weighting method21,38. The subjective weighting method is represented by the analytic hierarchy process (AHP), Delphi method, and so on. It has the advantage of simplicity, but the disadvantage is too subjective and randomness because it was completely dependent on the knowledge and experience of decision makers. The objective weighting method is represented by the entropy weighting method (EWM), principal component analysis, variation coefficient method, and so on. And it has been widely recognized for reflecting the variability of assessment results18, but the values of indexes have significant influence and the calculation results are not stable. Considering the limitations of the single weighting method, the weights of each assessment index in this study were determined by the combination of subjective weight and objective weight. Among them, the subjective weighting selected the AHP, and the objective weighting selected the EWM (Table 2). The formula is as follows:$$ w_{{j}} = alpha w_{{j}}^{{{AHP}}} + (1 – alpha )w_{{j}}^{{{EWM}}} $$
(10)
$$ w_{{j}}^{{{EWM}}} = d_{{j}} /sumlimits_{i = 1}^{m} {d_{{j}} } $$
(11)
$$ d_{{j}} = 1 – e_{{j}} $$
(12)
$$ e_{{j}} = – ksumlimits_{i = 1}^{n} {f_{{{ij}}} ln (f_{{{ij}}} )} ,;k = 1/ln (n) $$
(13)
$$ f_{{{ij}}} = X^{prime}_{{{ij}}} /sumlimits_{i = 1}^{n} {X^{prime}_{{{ij}}} } $$
(14)
where (W_{{j}}^{{}}) is the combined weight. (W_{{j}}^{{_{AHP} }}) is the weight of the j-th index of the AHP, (W_{{j}}^{{{EWM}}}) is the weight of the j-th index of the EWM, dj is the information entropy of the j-th index, ej is the entropy value of the j-th index, (f_{{{ij}}}) is the proportion of the index value of the j-th sample under the i-th indexm, (X^{prime}_{{{ij}}}) is the standardized value of the i-th sample of the j-th index, m is the number of index, n is the number of samples, and (alpha) was taken as 0.5.Table 2 Weight of assessment index.Full size tableSince the dimensions of indexes are different, it is necessary to unify the dimensions of the index to avoid the errors caused by direct calculation to make the evaluation results inaccurate. The range standardization was used to normalize the index data and bound its value in the interval [0, 1], the range standardization can be expressed as follows15,23:$$ {text{Positive indicator}}left( + right):A_{{{ij}}} = (X_{{{ij}}} – X_{{{jmin}}} )/(X_{{{jmax}}} – X_{{{jmin}}} ) $$
(15)
$$ {text{Negative indicator}}left( – right):A_{{{ij}}} = (X_{{{jmax}}} – X_{ij} )/(X_{{{jmax}}} – X_{{{jmin}}} ) $$
(16)
Additionally, we divided the LEH index into five levels from high to low using an equal-interval approach as follows40: [1–0.8) healthy, [0.8–0.6) sub-healthy, [0.6–0.4) moderately healthy, [0.4–0.2) unhealthy, [0.2–0] pathological, corresponding level I–V. And the level transfer of LEH in different periods was divided into three types: improvement type, degradation type, and stabilization type. For example, III-II means that the transfer from level III to level II is the improvement type.Spatial autocorrelation analysisSpatial autocorrelation analysis is one of the basic methods in theoretical geography. It could deeply investigate the spatial correlation characteristics of data, including global spatial autocorrelation and local spatial autocorrelation23. The global spatial autocorrelation uses global Moran’s I to evaluate the degree of their spatial agglomeration or differentiation of an attribute value in the study area. The local spatial autocorrelation is a decomposed form of the global spatial autocorrelation18,21, including four types: HH(High-High), LL(Low-Low), HL(High-Low), LH(Low–High). In this study, spatial autocorrelation analysis was applied to study the spatial correlation characteristics of blue–green space LEH. The calculation formulas are as follows:$$ I = frac{{Nsumlimits_{i} {sumlimits_{v} {W_{iv} (Y_{i} – overline{Y} )(Y_{v} – overline{Y} )} } }}{{(sumlimits_{i} {sumlimits_{v} {W_{iv} } } )sumlimits_{i} {(Y_{i} – overline{Y} )} }} $$
(17)
$$ I_{i} = frac{{Y_{i} – overline{Y} }}{{S_{x}^{2} }}sumlimits_{v} {left[ {W_{iv} (Y_{i} – overline{Y} )} right]} $$
(18)
where N is the number of space units, (W_{iv}) is the spatial weight, (Y_{i} ,Y_{v}) are the variable attribute values of the area (i,v), (overline{Y}) is the variable mean, (S_{x}^{2}) is the variance, (I) is the global Moran’s I index, and (I_{i}) is the local Moran’s I index. More
