EN中文

Journal of Geographical Sciences >

2022 , Vol. 32 >Issue 11: 2251 - 2270

DOI: https://doi.org/10.1007/s11442-022-2046-x

Research Articles

Urban construction land demand prediction and spatial pattern simulation under carbon peak and neutrality goals: A case study of Guangzhou, China

  • HU Xintao , 1, 2 ,
  • LI Zhihui 1, 3 ,
  • CAI Yumei 4 ,
  • WU Feng , 1, 3, *
Expand
  • 1. Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China
  • 2. College of Land Science and Technology, China Agricultural University, Beijing 100193, China
  • 3. University of Chinese Academy of Sciences, Beijing 100049, China
  • 4. Research Centre for Territorial Spatial Planning, Ministry of Natural Resources of the People’s Republic of China, Beijing 100812, China
*Wu Feng (1979‒), Associate Professor, specialized in complex system modeling and sustainable development. E-mail:

Hu Xintao (1997‒), PhD Candidate, specialized in econometric and policy evaluation research. E-mail:

Received date: 2021-11-03

  Accepted date: 2022-03-14

  Online published: 2022-11-25

Supported by

National Natural Science Foundation of China(41971233)

Fold

Abstract

Urban construction land has relatively high human activity and high carbon emissions. Research on urban construction land prediction under carbon peak and neutrality goals (hereafter “dual carbon” goals) is important for territorial spatial planning. This study analyzed quantitative relationships between carbon emissions and urban construction land, and then modified the construction land demand prediction model. Thereafter, an integrated model for urban construction land demand prediction and spatial pattern simulation under “dual carbon” goals was developed, where urban construction land suitability was modified based on carbon source and sink capacity of different land-use types. Using Guangzhou as a case study, the integrated model was validated and applied to simulate the spatiotemporal dynamics of its urban construction land during 2030-2060 under baseline development and “dual carbon” goals scenarios. The simulation results showed that Guangzhou’s urban construction land expanded rapidly until 2030, with the spatial pattern not showing an intensive development trend. Guangzhou’s urban construction land expansion slowed during 2030-2060, with an average annual growth rate of 0.2%, and a centralized spatial pattern trend. Under the “dual carbon” goal scenario, Guangzhou’s urban construction land evolved into a polycentric development pattern in 2030. Compared with the baseline development scenario, urban construction land expansion in Guangzhou during 2030-2060 is slower, with an average annual growth rate of only 0.1%, and the polycentric development pattern of urban construction land was more prominent. Furthermore, land maintenance and growth, that is, a carbon sink, is more obvious under the “dual carbon” goals scenario, with the forest land area nearly 10.6% higher than that under the baseline development scenario. The study of urban construction land demand prediction and spatial pattern simulation under “dual carbon” goals provides a scientific decision-making support tool for territorial spatial planning, aiding in quantifying territorial spatial planning.

Key words： carbon peak and neutrality goals; urban construction land; demand prediction; spatial pattern simulation; Guangzhou

Cite this article

HU Xintao , LI Zhihui , CAI Yumei , WU Feng . Urban construction land demand prediction and spatial pattern simulation under carbon peak and neutrality goals: A case study of Guangzhou, China[J]. Journal of Geographical Sciences, 2022 , 32(11) : 2251 -2270 . DOI: 10.1007/s11442-022-2046-x

1 Introduction

High-quality development is the coordinated development of humans and nature, which improves resource efficiency and avoids negative environmental effects. It realizes the goal of ecological-economic coordinated development by constantly regulating human behavior. Human activities are a primary cause of land use and land cover changes (LULCC) that affect the natural environment (Jin et al., 2020). As the main gathering area for human activities, urban construction land is an important carrier of environmental regulation (Mwangi et al., 2016). From 1870 to the present, LULCC’s cumulative carbon emissions accounted for approximately 33% of the total anthropogenic carbon emissions, of which urban development is a significant contributor (Peng et al., 2017). During urbanization, land resources utilization in some cities is inefficient because of the disorganized development pattern of the regional economy and disordered expansion of urban space (Lau et al., 2005). Climate change pressure and the responsibility to reduce emissions are driving cities towards a green, low-carbon, and high-quality development mode (Jin et al., 2019). China’s “dual carbon” goals are a major strategic decision showing its resolve to address climate change, and will profoundly impact the expansion rate and spatial development trend of urban construction land (Gong et al., 2021). Under the “dual carbon” goals, energy and industrial structures will be further optimized, new industries and technologies will be developed more rapidly, the intensive development of construction land will become the norm, and ecological protection policies, such as forest and grassland protection and arable land protection will be further strengthened (Li and Huang, 2021). Therefore, the analysis of future spatiotemporal variation and trends of urban construction land demand and intensive use in cities under the “dual carbon” goals can provide a basis for optimal land use management and sustainable development in the future and is fundamental for achieving the “dual carbon” goals.
Multi-scale and -level land-use demand prediction and simulation models are scientific tools for studying the evolution of urban construction land patterns. The main methods for urban construction land demand prediction include driving force analysis, trend extrapolation, and stochastic process analysis. For example, Samie et al. (2017) considered the natural (temperature, precipitation, etc.) and socioeconomic (population, GDP, etc.) driving factors of urban construction land evolution and used multiple regression models to predict and simulate land use in Punjab Province, Pakistan. However, this method can only be used for short-term land use demand prediction. Zhang et al. (2013) used the grey prediction model GM (1,1) to predict the land-use area of Wu’an City, China. Sun et al. (2016) calculated the land-use transfer probability using a Markov model based on the land-use grid data of 1989, 2000, and 2010, on which the land-use area of Zhangjiakou City was predicted. The prediction results of the above models strictly follow the historical development inertia. However, the future land demand for urban development in the new era needs to consider the “dual carbon” goals, and the existing prediction and analysis methods of urban construction land have not yet coupled the quantitative relationship between carbon emissions and urban construction land.
With the enrichment of remote sensing inversion data of fine-scale urban patterns, the simulation scale of the spatial pattern of urban construction land has also been refined. Cellular automata (CA)-based land use simulation models are mainstream research methods for urban construction land spatial pattern simulation. On one hand, it can simulate complex land-use pattern changes through simple transformation rules; on the other hand, it can be coupled with remote sensing and geographic information data to simulate land use pattern changes under different driving factors (Feng et al., 2018). Land use simulation models developed based on CA include the CLUE-S model (Verburg et al., 2002) which is suitable for small- and medium-scale simulations with high-resolution spatial data; the GWR-CA model (Feng and Tong, 2018) which combines CA with geo-weighted regression to show the impact of driving forces on land-use change; the FLUS model, which combines CA with system dynamics to consider the impact of climate change on land-use change; the CFLUS model, which simulates regional urban development; the OS-CA model, which simulates the growth of open space (parks, squares, etc.); the PLUS model, which simulates the generation and development of arbitrary multiple land-use types; and the MCCA model, which simulates land competition at subcellular scales (Liu et al., 2017; Liang et al., 2020; Liang et al., 2021). With the deepening of the “dual carbon” goals, intensive development patterns and low-carbon restructuring will become new trends in urban construction land evolution. Urban land has been shown to have the most concentrated and intense carbon emissions (Chuai et al., 2015). Globally, annual carbon emissions from urban land due to energy consumption, economic growth, and transport exceed 70% of the overall carbon emissions (Song et al., 2015; Li et al., 2018). Therefore, trends in carbon emissions are an important aspect of future trends in urban development (Li et al., 2020). Existing research methods mainly focus on improving the research fineness and refining the calculation of the probability of raster conversion based on the analysis of driving factors to improve model simulation accuracy, whereas relatively few studies consider the simulation of the evolution of urban construction land patterns under “dual carbon” goals.
The measurement of the contribution of urban size, structure, and pattern to carbon emissions from human activities, and prediction and simulatation of urban construction land spatiotemporal patterns in the future under the “dual carbon” goals scenario is necessary for quantitative and low-carbon development of territorial spatial planning. Therefore, this research attempts to establish an integrated simulation method for urban construction land demand prediction and spatial pattern simulation under the “dual carbon” goals. Using Guangzhou as an example, we simulated and compared the differences in the evolution of urban construction land demand and spatial patterns in 2030 and 2060 under two scenarios of baseline development and “dual carbon” goals, clarifying the impact of the “dual carbon” goals on future land use changes. Furthermore, by comparing the model results with the development plan of Guangzhou, the rationality and accuracy of our model, which simulates urban development and land use changes under the constraint of carbon emissions goals, were validated, providing a baseline for future urban territorial spatial planning and optimal management.

2 Data and methods

2.1 Study area

Guangzhou is the capital city of Guangdong Province and the most developed coastal region in Southeast China. It is located in the central and southern parts of Guangdong Province. It borders the South China Sea and is close to the mouth of the lower reaches of the Pearl River Basin. It spans between 112°57°E-114°03°E and 22°26°N-23°56°N, with a total area of approximately 7434.4 km2 (Figure 1). The topography of Guangzhou is diverse, with the terrain being high in the northeast and low in the southwest. Northeast Guangzhou is a low and medium mountainous area dominated by forest land, the central part of which is a hilly basin, and the south is a coastal alluvial plain. The Statistical Bulletin of the National Economic and Social Development of Guangzhou in 2020 shows that Guangzhou achieved a regional GDP of more than 2.50 trillion yuan (RMB) 2020, with a resident population of 18.68 million. Its urbanization rate was 86.46% in 2020, an increase of over 25% compared to 2005. According to the Emissions Database for Global Atmospheric Research (EDGAR), Guangzhou’s urban carbon emissions have been growing rapidly over the past two decades, with carbon emissions exceeding 65 million tons in 2018. Guangzhou has the highest per capita GDP and per capita energy consumption, which is higher than that of Beijing and Shanghai (Wang et al., 2019). Guangzhou City is a pilot zone for low-carbon cities, as announced by China’s National Development and Reform Commission (NDRC) in 2012 (Wang et al., 2015). Therefore, taking Guangzhou as the study area for this research is significant for the construction of a low carbon city in China and the achievement of the “dual carbon” goals.
Figure 1 Geographical location of Guangzhou

2.2 Data

In this study, the socio-economic and carbon emissions time series data of Guangzhou, as well as the traffic network, dual evaluation suitability (DES), carbon emissions, and carbon sink suitability (CECSS), and main functional area (MFA) raster data were used. Data on the population density, per capita GDP, energy consumption per unit of GDP, gross industrial product value, and urban construction land area of Guangzhou were obtained from the Guangzhou Statistical Yearbook (http://tjj.gz.gov.cn/) (Table 1). Land use data were obtained from the Second Round National Land Survey and Annual National Land Use Change Survey data. The 100 m grid data of land use in Guangzhou in 2012, 2015, and 2018 were selected and classified into six types: arable land, forest land, grassland, water area, urban construction land, and unused land. Vector data for major roads, railways, and highways in Guangzhou were obtained from the Data Center for Resources and Environmental Sciences (RESDC) of the Chinese Academy of Sciences. The grid data for the evaluation of the resources and environment carrying capacity, territorial space suitability, and main function zoning of Guangzhou were obtained from the Chinese Land Surveying and Planning Institute. Urban carbon emissions data of the year 1970-2018 were obtained from EDGAR (https://ec.europa.eu/info/index).
Table 1 Descriptive statistics of the indicators
Variables Min Max Mean Variance
Urban construction land area (km2) 735 1300 1017 31,200
Resident population density (person/km2) 1227 2005 1639 62,289
GDP per capita (yuan/person) 54,627 153,373 104,038 1,079,426,922
Energy consumption per unit of GDP (million tons/billion dollars) 27 78 44 256
Proportion of added value of the tertiary industry in GDP (%) 58 70 63 16

2.3 Methods

To simulate the evolution of urban construction land patterns under the “dual carbon” goals, it is necessary to determine the relationship between carbon emissions and urban development at the present stage and to analyze the change rules of influencing factors of urban development under the “dual carbon” goals. Therefore, we established an integrated model for predicting urban construction land demand and simulating the spatial pattern under the “dual carbon” goals (Figure 2). First, the model uses the Stochastic Impacts by Regression on Population, Affluence, and Technology (STRIPAT) econometric model to clarify the quantitative relationship between carbon emissions and urban construction land based on historical data. Then, the future demand for urban construction land is primarily predicted using the forecasted carbon emissions estimated by applying the neural network autoregressive (NNAR) model. This is assumed to result in the baseline development scenario. Then, the specific future demand for urban construction land, and its relevant influencing analysis modules and parameters under the “dual carbon” goals are further adjusted and validated considering industrial structure, energy consumption intensity, land use carbon emissions, and carbon sink. Finally, the simulation of the spatial pattern of future urban construction land evolution under the baseline development and “dual carbon” goals scenarios are obtained through the Policy and Land Use Simulator for Carbon Emissions (PoLUS-C) model.
Figure 2 Framework for the integrated model for future urban construction land demand prediction and spatial pattern simulation under the “dual carbon” goals scenario

2.3.1 STIRPAT model

The STIRPAT model is an effective method to analyze the relationship between urban construction land and carbon emissions. It is derived from the impact, population, affluence, and technology (IPAT) model, which is a widely accepted environmental stress formula for analyzing the impact of human activities on the environment. In the early 1970s, IPAT emerged from discussions about the driving forces of human activities on the environment. It is believed that the impact of human activities on the environment comes from population (P), affluence (A), and technology (T); thus, the impact was defined as the product of P, A, and T, i.e., I = PAT. Population is mainly expressed by population density, affluence by GDP per capita, and technology by the impact formed per unit of GDP. The traditional IPAT model assumes that “all influences affect environmental stress in equal proportions”. York et al. (2003) extended the IPAT model to the STIRPAT model, which is a stochastic model that proposes a stochastic influence model for population, affluence, and technology. The STIRPAT model is expressed as follows:
$I = aPbAcTdε$
where a is a constant, b, c and d are indices of population, affluence, and technology drivers, respectively, and ε is the error term. Unlike IPAT, the STIRPAT model can be used for empirical hypothesis testing (Zhang and Zhao, 2019; Xu et al., 2020). To facilitate estimation and hypothesis testing, Eq. (1) is usually transformed into a logarithmic form, as shown in Eq. (2), where α and φ are the logarithmic values of a and ε respectively. b, c, and d denote the elasticity coefficients of carbon emissions on population, affluence, and technology, respectively, which means that every 1% change in population, affluence, and technology causes environmental changes of b%, c%, and d%, respectively.
$lnI=\alpha +blnP+clnA+dlnT+\varphi$
In this study, we used the extended STIRPAT model to establish an analytical model to clarify the relationship between urban construction land and carbon emissions. Based on existing studies, carbon emissions (C) were taken as the dependent variable and urban construction land area (U), population density (R), GDP per capita (G), energy consumption per unit of GDP (E), the proportion of the added value of the tertiary industry in GDP (T), and one-period-lagged carbon emissions (Clag) were taken as the driving factors of carbon emissions. Before the regression analysis, a principal component analysis was conducted to reduce the influence of the correlation between independent variables on the measurement results, which enables the results to be predicted in the future (Chun and Keles, 2010). Then, the coefficients are solved by the partial least squares regression method, and the variance of the regression coefficients is estimated using the jackknife method to calculate the significance of the regression coefficients (Eq. 3).
$lnC=\alpha +blnU+clnR+dlnG+elnE+flnT+gln{{C}_{lag}}+\varphi $

2.3.2 NNAR model

The autoregressive model is an effective method for simulating carbon emissions patterns and predicting changes in carbon emissions. However, the NNAR model that combines different models is more suitable for nonlinear and nonstationary time series simulations (Ji and Chee, 2011). The NNAR model is similar to the linear autoregressive model, which uses the lagged value as an independent variable. The time-series data and lag value of the time series can also be used as inputs to the neural network. Similarly to the linear autoregressive model that uses the lagged value as an independent variable, the time-series variable and the lagged value of the time-series variable can also be used as the input of the NNAR model. In this study, C(t-1), C(t-2),..., C(t-d) are the lagged terms for carbon emissions, and d is the lagged time parameter. In the NNAR model, the network was created and trained in an open loop using real values as response variables, making the training results closer to the real values. The network is converted to a closed-loop at the end of training and the predicted values are used to provide new response inputs to the network and applied to a discrete non-linear autoregressive neural network, and a discrete non-linear autoregressive model of carbon emissions is described (Eq. 4).
${{C}_{t}}=h\left( {{C}_{t-1}},{{C}_{t-2}},\ldots,{{C}_{t-d}} \right)+{{\varepsilon }_{t}}$
Neural network training aims to approximate function h(•) by optimizing the network weights and neuron biases. Thus, the NNAR model is defined as follows:
${{C}_{t}}={{\alpha }_{0}}+\underset{j=1}{\overset{k}{\mathop \sum }}\,{{\alpha }_{j}}\varphi \left( \underset{j=1}{\overset{k}{\mathop \sum }}\,{{\beta }_{ij}}{{C}_{t-i}}+{{\beta }_{0j}} \right)+{{\varepsilon }_{t}}$
where i is the number of input variables, and k is the number of hidden layers with an activation function. βij is the parameter corresponding to the weight of the connection between the hidden layers, such as input unit i and hidden unit j; αj is the weight of the connection between hidden unit j and output unit, β0j and α0 are the constant terms of the hidden unit j and output unit, respectively.

2.3.3 PoLUS-C model

Based on the analysis of the factors influencing the evolution of the spatial pattern of urban construction land under the “dual carbon” goals, the PoLUS-C model based on CA (meta-cellular automata) was established (Figure 3). The CA model combines the mathematical theory of automata’s self-replication and randomness with the geographic information system (GIS) spatial grid map, which can effectively reflect the relationships between different land grids (Tobler, 1979). Combined with existing research, we established five driving force modules of neighborhood influence (N), accessibility (A), suitability (S), main functional area (MFA), and randomness (v), which were multiplied to obtain the development potential index of urban construction land on each grid (Eq. 6).
$P_{j,c}^{t}=N_{j,c}^{t}A_{j,c}^{t}S_{j,c}^{t}MFA_{j,c}^{t}v_{j,c}^{t}$
Figure 3 PoLUS-C model framework
The neighborhood influence (N) module reflects the ability of urban construction land to encroach on other land-use types during the historical period and indicates the extension of historical development inertia. This module requires setting the influence range and influence value. The influence range represents the distance that the impact of a grid can reach, and the influence value represents the attractiveness of a grid to make its surrounding grids convert their land use types to the same type. PoLUS-C allows setting influence values for 1-5 grid distances on either side of the central grid, with diagonal values calculated using Pythagoras’ theorem.
The development of a traffic network is an important constraint factor in the evolution of spatial patterns of urban construction land. The accessibility (A) module reflects the attractiveness of the traffic network to the development of urban construction land, which is represented by the distance of the main roads (national, provincial, and country roads), railways, and highways to each grid (Eq.7).
${{a}_{j}}=1+{{\left( \frac{D}{{{\delta }_{j}}} \right)}^{-1}}$
where a is the accessibility of the grid, D is the Euclidean distance from each grid to the nearest traffic network, and δj is the accessibility coefficient of the different traffic networks.
Under the “dual carbon” goals, the capacity of carbon emissions and carbon sinks of land use have become important influencing factors for urban construction land planning and development. Based on the existing research on the carbon emissions and carbon sink capacity of land use in China, the net carbon emissions per unit for different land types were normalized and used as the basis for the establishment of carbon emissions and carbon sink suitability (CECSS). Then, the suitability assessment (S) module for urban construction land was constructed based on CECSS and also considered locational conditions, resources, environment carrying capacity, and territorial space development suitability (DES).
The MFA is an important baseline for future urban development, which determines the core functions of each geographical unit at different scales, and is a general layout map for future territorial spatial planning (Wang and Lu, 2014). The MFA is used as one of the modules that affect the simulation of the spatiotemporal evolution of urban construction land. The randomness (v) module represents the unconsidered uncertainties, which are calculated by a random number function, where ran is a random number that follows a uniform distribution within the range of 0-1 and α is the random effect size (Eq. 8).
$v_{j,c}^{t}={{\left( -\ln \left( 1-ran \right) \right)}^{\alpha }}$

3 Results

3.1 Impacts of urban construction land on carbon emissions in Guangzhou

The standardized coefficients of the established STRIPAT model were solved, and the significance of the regression coefficients was obtained. The P-values of each variable were less than 0.001, indicating the significant impact of the driving factors on carbon emissions (Table 2). The elastic coefficients of carbon emissions for each variable in the STRIPAT model were then obtained based on the relationship between the coefficients of the original variables and the standardized coefficients (Eq. 9). The results show that the elastic coefficient of urban construction land is 0.1, which means that every 1% increase in urban construction land leads to a 0.1% increase in carbon emissions in Guangzhou. These results are consistent with the results of existing studies. The relationship between economic development, energy consumption, and carbon emissions in China presents an environmental Kuznets curve; that is, with economic and energy consumption development, carbon emissions show an inverted U-shaped curve that first rises and then declines (Shen and Sun, 2016; Lv et al., 2021). The fitted curve of carbon emissions in Guangzhou changed from lower than actual carbon emissions to higher than actual carbon emissions during 2009-2018, indicating that Guangzhou’s carbon emissions are at the inflection point of the environmental Kuznets curve (Figure 4). Under the “dual carbon” goals, optimizing industrial structure, reducing energy consumption intensity, and promoting intensive development are the main directions for urban construction land in Guangzhou in the future. Therefore, the demand for urban construction land in Guangzhou under the “dual carbon” goals should be adjusted based on the above model results.
$C=a{{U}^{0.1}}{{R}^{0.12}}{{G}^{0.057}}{{E}^{-0.056}}{{T}^{0.28}}C_{lag}^{0.11}$
Table 2 Standardized regression coefficients and significance tests of the STIRPAT model
Variables Estimate Std. Error t value Pr (>|t|)
Urban construction land area (km2) 0.14 0.03 5.96 0.00005***
Resident population density (person/km2) 0.16 0.04 5.07 0.0002***
GDP per capita (yuan/person) 0.16 0.04 5.25 0.0001***
Energy consumption per unit GDP (million tons/billion dollars) -0.15 0.04 -5.34 0.0001***
Proportion of added value of the tertiary industry in GDP (%) 0.14 0.03 4.45 0.0007***
One-period-lagged carbon emissions (million tons) 0.14 0.03 4.41 0.0007***
Figure 4 Fitted results for standardized carbon emissions in Guangzhou during 2005-2018

3.2 Predictions of carbon emissions in Guangzhou

The urban carbon emissions of Guangzhou for 2019-2060 were predicted based on the NNAR model using 1970-2018 urban carbon emissions data (Figure 5). The results showed that the predicted urban carbon emissions of Guangzhou in 2030 is 69.8 million tons, with an increase of 5.36% compared with that in 2018. The predicted carbon emissions of Guangzhou in 2060 are 70.2 million tons, which is the same as that in 2030. Combined with the results of the carbon emissions effect of urban construction land in Guangzhou, the area of urban construction land in Guangzhou was estimated to expand by 53.6% and 6% during 2018-2030 and 2030-2060, respectively. Therefore, the development of urban construction land in Guangzhou during 2018-2030 was still dominated by expansion, although considering internal pattern optimization and intensive development, it is mainly dominated by internal spatial pattern optimization during 2030-2060.
Figure 5 Predicted results of urban carbon emissions in Guangzhou (red dotted line is the predicted results of urban carbon emissions during 2019-2030, the blue line is the predicted results of urban carbon emissions during 2030-2060, 80%, and 95% prediction intervals are calculated by bootstrap method)
The above results can only reflect changes in urban construction land demand and spatial pattern evolution in Guangzhou under inertia development. Under the “dual carbon” goals, the upgrading and adjustment of industrial structures in Guangzhou will change the increasing trend of current urban construction land demand, and stricter environmental protection policies will force the internal optimization process of urban construction land, thus enhancing the carbon sink function of land use (Lin and Xia, 2013). Therefore, the factors influencing the urban construction land demand and pattern evolution in Guangzhou under the “dual carbon” goals need to be further determined.

3.3 Urban construction land demand and pattern evolution parameters setting in Guangzhou

Economic development and energy consumption are the main drivers of carbon emissions (Ma et al., 2019). The predicted results of Guangzhou’s urban construction land demand and parameters for its spatial pattern simulation were further calibrated based on the analysis of the development laws of Guangzhou’s urban construction land, carbon emissions, industrial structure, and energy intensity. Subsequently, the urban construction land demand and spatial pattern evolution parameter schemes under the two scenarios of baseline development and “dual carbon” goals were obtained for Guangzhou’s future urban construction land simulation during 2030-2060.
Under the “dual carbon” goals, protecting the beauty of ecological space, ensuring the quality and efficiency of agricultural space, and promoting the compact and intensive urban space are important measures for the territorial spatial planning of Guangzhou to promote the construction of ecological civilization. As a city with the earliest reform and fastest economic development in China, Guangzhou represents rapid urbanization (Yang et al., 2020). Land use changes in the region are significant, the contradiction between urban development and the ecosystem is increasingly prominent, and ecological risks continue to intensify (Guo et al., 2021). Land use and land cover changes caused by urban expansion are some of the main reasons for the increase in carbon emissions. Rapid urbanization has greatly changed Guangzhou landscape patterns, with a large area of forest land and reduced arable land (Xu et al., 2016). According to the second national land survey, construction land in Guangzhou expanded by 11,402 ha during 2012-2018, whereas the areas of forest and arable land decreased by 5850 and 4453 ha, respectively (Figure 6). Therefore, under the “dual carbon” goals, the carbon emissions and sink capacity of land use is an important influencing factor for the evolution of the spatial pattern of urban construction land.
Figure 6 Area changes of different land-use types in Guangzhou during 2012-2018
Reducing the proportion of the secondary industry is an effective means of reducing the overall carbon emissions intensity. Under the “dual carbon” goals, low carbon industries such as the financial industry, logistics industry, information service industry in Guangzhou will be further developed, and the original high energy consumption industries such as steel and cement plants will be gradually upgraded or replaced. Guangzhou’s economic development is mainly based on secondary and tertiary industries. With the continuous development of a low-carbon economy, the proportion of the secondary industry has decreased from 41% in 2000 to 27% in 2018, and its energy consumption and carbon emissions are also decreasing annually. The tertiary industry has been developing policy support and its proportion has also been increasing, rising from 55% in 2000 to 71% in 2018. Therefore, the growth of urban construction land demand in Guangzhou will gradually slow down under the “dual carbon” goals, which should be lower than the result under the baseline development scenario. Energy consumption and intensity are also the main drivers of carbon emissions. Recently, Guangzhou has strengthened the review of energy conservation of fixed-asset investment projects and promoted the process of clean energy projects, such as photovoltaic power generation, which has ensured stable economic development while continuously reducing energy consumption intensity and improving energy consumption efficiency (Figure 7). Under the “dual carbon” goals, strengthening energy control and promoting cleaner production in the industrial sectors is an important tool for energy conservation and emissions reduction in Guangzhou (Jiao et al., 2019). Therefore, Guangzhou will slow down the expansion of construction land, further optimize the layout of construction land, promote new energy projects, and reduce energy consumption intensity.
Figure 7 Change in energy consumption intensity in Guangzhou
Based on the above analysis, we attempted to parameterize the impact of the “dual carbon” goals on urban construction land development in the demand and spatial pattern modules, forming a parametric scheme for the simulation of future urban construction land in Guangzhou for 2030-2060 under two scenarios of baseline development and “dual carbon” goals, thus to obtain a more intuitive comparison of the impact of the “dual carbon” goals on future urban construction land development (Table 3). The urban construction land demand under the baseline development scenario was derived according to the above STRIPAT and NNAR results; that is, the urban construction land area in Guangzhou will increase by 100,298 ha (53.6%) and 17,244 ha (6.0%) during 2018-2030 and 2030-2060 respectively, with average annual growth rates of 4.46% and 0.5%. The demand for urban construction land under the “dual carbon” goals scenario will be reduced by 0.94% and 0.25%, respectively, based on the average annual growth, that is, the area of urban construction land in Guangzhou will increase by 78,867 ha (42.2%) and 7980 ha (3.0%) during 2018-2030 and 2030-2060, respectively, with an average annual growth rate of 3.51% and 0.25%, respectively. In addition, the weights for modules in the PoLUS-C model for Guangzhou’s urban construction land spatial pattern simulation under the baseline development scenario were derived from the existing literature. Compared to the baseline development scenario, the weights under the “dual carbon” goals scenario were adjusted from the baseline development scenario based on the above analysis. The CECSS factor was added to the suitability module to consider the carbon emissions and sink capacity of land use, and the weighting of this factor was set to increase for the period 2030-2060.
Table 3 Parameter schemes for urban construction land demand prediction and spatial pattern simulation in Guangzhou (the unit for urban construction land demand ha, other parameters are the weighting between modules affecting the evolution of urban construction land spatial pattern)
Parameter type Baseline development scenario “Dual carbon” scenario
2030 2060 2030 2060
Urban construction land demand (ha) 287,420 304,665 265,990 273,969
Neighborhood 20 20 20 20
Accessibility Main roads 5 5 5 5
Railroad 10 10 10 10
Highway 5 5 5 5
Suitability DES 20 20 10 7
CECSS / / 10 14
MFA 20 20 20 20

3.4 Urban construction land demand prediction and spatial pattern simulation in Guangzhou

The neighborhood module parameters in the PoLUS-C model were calibrated according to the changes in land-use patterns caused by the expansion and contraction of urban construction land during the historical period. Land use data from the Second Round National Land Survey and Annual National Land Use Change Survey showed that the expansion area of urban construction land in Guangzhou during 2012-2015 was much larger than the reduction area. This expansion was mainly based on the occupation of forest land and arable land. The areas of these two land use types decreased by 3048 and 1995 ha, respectively (Figure 8). The accuracy of the PoLUS-C model was verified by simulating the land use pattern in Guangzhou in 2018, where the Kappa coefficient was 0.97, and the overall accuracy of the model was verified to be 0.99.
Figure 8 Urban construction land expansion and reduction in Guangzhou during 2012-2015

3.4.1 Simulation results of urban construction land in Guangzhou in 2030

Under the baseline development scenario, the area of urban construction land in Guangzhou in 2030 is 287,421 ha, an increase of 53.6% from 2018, which is slightly higher than the growth rate presented in the Guangzhou Urban Master Plan (2017-2035). Regarding regional distribution, urban construction land development in Guangzhou is mainly in the central and southern regions, including Nansha, Panyu, Tianhe, Yuexiu, Liwan, and most of the Haizhu districts, as well as the southern parts of Huadu and Zengcheng districts and the western part of Baiyun District. However, the growth of urban construction land was less in the ecological conservation zones north of Guangzhou (Figure 9). Regarding land-use change, the expansion of urban construction land in Guangzhou is still dominated by the occupation of forest land and water areas, which decreased by 59,360 and 36,582 ha, respectively (Table 4).
Figure 9 Simulation results of the urban construction land spatial pattern in Guangzhou in 2030 (a. Baseline development scenario; b. “Dual carbon” goals scenario)
Table 4 Comparison of simulation results by land-use type under different scenarios in Guangzhou in 2030
Land-use type Area (ha)
Baseline development scenario Area changes compared to 2018 “Dual carbon” goals scenario Area changes
compared to 2018
Arable land 81,169 -1233 70,578 ‒11,824
Forest land 297,947 -59,360 328,294 -29,013
Grassland 672 -1955 921 -1706
Water area 58,624 -36,582 59,975 -35,231
Unused land 471 -1209 546 -1134
Urban construction land 287,421 100,298 265,990 78,867
Under the “dual carbon” goals scenario, the area of urban construction land in Guangzhou in 2030 is 265,990 ha, increasing by 42.15% since 2018, which is generally consistent with the growth rate in the Guangzhou Urban Master Plan (2017-2035). Regarding regional distribution, urban construction land development in Guangzhou was still mainly in the central and southern regions, which is consistent with the results under the baseline development scenario (Figure 9). Regarding land-use change, the expansion of the urban construction land of Guangzhou is more evenly distributed, with forest land, arable land, and water area decreasing by 29,013, 11,824, and 35,231 ha, respectively (Table 4). In comparison, the spatial pattern of urban construction land in Guangzhou forms an intensive polycentric development pattern, with a more obvious effect of forest and grassland protection under the “dual carbon” goals scenario.

3.4.2 Simulation results of urban construction land in Guangzhou in 2060

Under the baseline development scenario, the area of urban construction land in Guangzhou in 2060 is 304,665 ha, an increase of 6% compared to the simulation results in 2030. Urban construction land development during this period mainly focused on optimizing the spatial pattern. Regarding regional distribution, urban construction land development becomes more concentrated during 2030-2060, with urban construction land in the eastern part of Zengcheng district and the southern part of Nansha district developing towards their central areas (Figure 10). Regarding land-use change, the arable land area increased by 21,611 ha during 2030-2060, which is largely consistent with the results of the red line delineation for arable land in the Guangzhou Urban Master Plan (2017-2035), but with a time lag. The forest land and water areas decreased significantly during this period, by 23,861 and 14,640 ha, respectively, but the reduction rate has slowed down significantly compared to that during 2018-2030 (Table 5).
Figure 10 Simulation results of the urban construction land spatial pattern in Guangzhou in 2060 (a. Baseline development scenario; b. “Dual carbon” goals scenario)
Table 5 Comparison of simulation results by land-use type under different scenarios in Guangzhou in 2060
Land-use type Area (ha)
Baseline development scenario Area changes compared to 2030 “Dual carbon”
scenario		 Area changes compared to 2030
Arable land 102,780 21,611 113,925 43,347
Forest land 274,086 -23,861 303,310 -24,984
Grassland 450 -222 334 -587
Water area 43,984 -14,640 34,500 -25,475
Unused land 339 -132 265 -281
Urban construction land 304,665 17,244 273,970 7980
Under the “dual carbon” goals scenario, the urban construction land area of Guangzhou in 2060 is 304,665 ha, with an increase of 3% compared to the simulation results in 2030, with a more noticeable effect of optimizing the spatial pattern. Regarding regional distribution, the urban construction land pattern evolved more intensively under this scenario, mainly in the southern part of Huadu District, the main urban areas, such as the Baiyun and Panyu districts, and the northern part of Nansha District (Figure 10). Regarding land-use change, the arable land area grew by 43,347 ha during 2030-2060, which is largely consistent with the results of the arable land red line delineation in the Guangzhou Urban Master Plan (2017-2035). The forest land area exceeds 300,000 ha, which is 10% higher than the results under the baseline development scenario, indicating that environmental protection is more effective under the “dual carbon” goals scenario. In comparison, the expansion rate of urban construction land under the “dual carbon” goals scenario is significantly lower during 2030-2060, and the spatial pattern of urban construction land development shows a trend of contraction, with higher conservation effects on arable land, forest land, and other high-carbon sink land-use types.

4 Discussion

The results showed that the “dual carbon” goals would lead to a more rational land use pattern, with the reduction of forest and grassland areas slowing significantly, and the effect of arable land protection becoming more obvious. Meanwhile, the “dual carbon” goals will have a notable impact on the evolution of urban construction land demand and spatial patterns, which indirectly affects the transformation of industrial structure and energy consumption, and ultimately affects the development of the spatial pattern of the territorial space. Compared with existing studies, the simulation results of urban construction land in Guangzhou in 2030 were generally consistent with the results of existing studies, while the simulation results for 2060 were somewhat different from the results of existing studies (Lin et al., 2020). Regarding land demand prediction, existing prediction methods, such as artificial neural networks and Markov, focus on extrapolation of future demand through historical trends, and the results are mainly linear development (Liu et al., 2017; Gong et al., 2018). In contrast, this study analyzes the relationship between carbon emissions and urban construction land, and forecasts the demand for urban construction land from the perspective of “dual carbon” goals. For land use spatial pattern simulation, existing studies have mainly studied the factors influencing the spatial pattern evolution of construction land from the perspectives of economic development, government planning, and geographical and climatic conditions. Therefore, this study analyzes the important impact of the “dual carbon” goals on its development. It has been found that the inertia of urban development affects the short-term development trend of urban construction land to a certain extent, whereas the “dual carbon” goals will gradually increase its influence in the longer term. Therefore, compared with existing studies, we established a quantitative relationship between carbon emissions and urban construction land, as well as an in-depth analysis of the impact of the “dual carbon” goals on the future development of urban construction land in Guangzhou in terms of economic structure, land use change, and energy consumption. We reduced the influence of historical development trends on the model simulation results and made the model’s prediction of future urban development and land use changes in Guangzhou more realistic. In addition, the model framework can generally be applied to other regions in China to simulate changes in urban construction land demand and patterns from the perspective of carbon emissions constraints, whereas the parameters of each module in different regions need to be adjusted according to their local conditions.
This study has some limitations; for example, it has not considered the impact of the “dual carbon” goals on different types of urban construction land, such as residential areas, industrial areas, factories, and mines. The model found that urban construction land development after 2030 mainly focuses on internal optimization. Therefore, it is important to consider the spatial evolution of different urban construction lands in the future. In addition, the land use data used in this study are the primary land use type data, which mainly reflect the overall development trend of urban construction land. Future research should focus on further subdividing urban construction land types to study the evolution of the internal spatial patterns of urban construction land.

5 Conclusion

The deadline set by China for the realization of “dual carbon” goals is much earlier than that set by developed countries, such as Europe and the US, so the speed of low-carbon transition needs to be faster. The study of the development law of urban construction land under the “dual carbon” goals is a necessary step in the new era of territorial spatial planning. In this study, using Guangzhou as the study area, the relationship between carbon emissions and urban construction land was clarified, and the influence of the “dual carbon” goals on the future development of urban construction land was analyzed. An integrated model of urban construction land prediction and spatial pattern simulation under the “dual carbon” goals was constructed. Then, using this model, the urban construction land demand and spatial pattern of Guangzhou in 2030 and 2060 were simulated under baseline development and the “dual carbon” goals scenarios and the impacts of “dual carbon” goals on the development of urban construction land in the future were analyzed. Through empirical research in this study, the following conclusions were reached: First, every 1% increase in urban construction land in Guangzhou over the recent ten years will lead to an average increase in urban carbon emissions of 0.1%. Second, Guangzhou’s urban carbon emissions will increase by 5.36% and 0.6% during 2018-2030 and 2030-2060, respectively, under baseline development. Third, comparing the simulation results of the two scenarios, it was found that under the “dual carbon” goals scenario, the urban construction land in Guangzhou will develop more slowly, and a polycentric urban development pattern will be formed in 2030, and the centralized development trend of urban construction land will be more obvious during 2030-2060; the change of land use is more realistic, with forest and grassland area decreasing at a considerably slower rate.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Commoner B, 1990. Making Peace with the Planet. New York: Pantheon.

[2]
Chun H, Keles S, 2010. Simultaneous dimension reduction and variable selection with sparse partial least squares. Journal of Royal Statistical Society B, 72(1): 3-25.

DOI

[3]
Chuai X, Huang X, Wang W, et al, 2015. Land use, total carbon emissions change and low carbon land management in Coastal Jiangsu, China. Journal of Cleaner Production, 103: 77-86.

DOI

[4]
Dietz T, Rosa E A, York R, 2007. Driving the human ecological footprint. Frontiers in Ecology and the Environment, 5(1): 13-18.

DOI

[5]
Dong F, Yu B, Hadachin T, et al, 2018. Drivers of carbon emission intensity change in China. Resources, Conservation and Recycling, 129: 187-201.

[6]
Feng Y, Tong X, 2018. Dynamic land use change simulation using cellular automata with spatially nonstationary transition rules. GIScience and Remote Sensing, 55(5): 678-698.

DOI

[7]
Gong J, Hu Z, Chen W, et al, 2018. Urban expansion dynamics and modes in metropolitan Guangzhou, China. Land Use Policy, 72: 100-109.

DOI

[8]
Gong W, Wang C, Fan Z, et al, 2021. Drivers of the peaking and decoupling between CO2 emissions and economic growth around 2030 in China. Environmental Science and Pollution Research, 29(3): 3864-3878.

DOI

[9]
Guo L, Xi X, Yang W, et al, 2021. Monitoring land use/cover change using remotely sensed data in Guangzhou of China. Sustainability, 13(5): 2944.

DOI

[10]
Haseeb M, Hassan S, Azam M, 2017. Rural-urban transformation, energy consumption, economic growth, and CO2 emissions using STRIPAT model for BRICS countries. Environmental Progress & Sustainable Energy, 36(2): 523-531.

[11]
Ji W, Chee K C, 2011. Prediction of hourly solar radiation using a novel hybrid model of ARMA and TDNN. Solar Energy, 85(5): 808-817.

DOI

[12]
Jiao Y, Su M, Ji C, et al, 2019. Cleaner production instruments assisting sustainable transition at urban scale: A case study of Dongguan, a typical manufacturing city in China. Journal of Cleaner Production, 210: 1449-1461.

DOI

[13]
Jin G, Chen K, Wang P, et al, 2019. Trade-offs in land-use competition and sustainable land development in the North China Plain. Technological Forecasting and Social Change, 141: 36-46.

DOI

[14]
Jin G, Shi X, He D, et al, 2020. Designing a spatial pattern to rebalance the orientation of development and protection in Wuhan. Journal of Geographical Sciences, 30(4): 569-582.

DOI

[15]
Lau S S Y, Giridharan R, Ganesan S, 2005. Multiple and intensive land use: Case studies in Hong Kong. Habitat International, 29(3): 527-546.

DOI

[16]
Li C, Li Y, Shi K, et al, 2020. A multiscale evaluation of the coupling relationship between urban land and carbon emissions: A case study of Chongqing, China. International Journal of Environmental Research and Public Health, 17(10): 3416.

DOI

[17]
Li J S, Zhou H W, Meng J, et al, 2018. Carbon emissions and their drivers for a typical urban economy from multiple perspectives: A case analysis for Beijing city. Applied Energy, 226: 1076-1086.

DOI

[18]
Li L, Zhu G, Wu D, et al, 2021. Land suitability assessment for supporting transport planning based on carrying capacity and construction demand. PloS One, 16(2): e0246374.

DOI

[19]
Li T, Huang Q, 2021. Green thermal power dispatching based on environmental information during the energy transition period. Journal of Cleaner Production, 320: 128843.

DOI

[20]
Liang X, Guan Q, Clarke K C, et al, 2021. Understanding the drivers of sustainable land expansion using a patch-generating land use simulation (PLUS) model: A case study in Wuhan, China. Computers, Environment and Urban Systems, 85: 101569.

DOI

[21]
Lin W, Sun Y, Nijhuis S, et al, 2020. Scenario-based flood risk assessment for urbanizing deltas using future land-use simulation (FLUS): Guangzhou Metropolitan Area as a case study. Science of the Total Environment, 739: 139899.

DOI

[22]
Lin Z, Xia B, 2013. Sustainability analysis of the urban ecosystem in Guangzhou city based on information entropy between 2004 and 2010. Journal of Geographical Sciences, 23(3): 417-435.

DOI

[23]
Liu X, Liang X, Li X, et al, 2017. A future land use simulation model (FLUS) for simulating multiple land use scenarios by coupling human and natural effects. Landscape and Urban Planning, 168: 94-116.

DOI

[24]
Lv D, Wang R, Zhang Y, 2021. Sustainability assessment based on integrating EKC with decoupling: Empirical evidence from China. Sustainability, 13(2): 655.

DOI

[25]
Ma X, Wang C, Dong B, et al, 2019. Carbon emissions from energy consumption in China: Its measurement and driving factors. Science of the Total Environment, 648: 1411-1420.

DOI

[26]
Mwangi H M, Julich S, Patil S D, et al, 2016. Relative contribution of land use change and climate variability on discharge of upper Mara River, Kenya. Journal of Hydrology: Regional Studies, 5: 244-260.

[27]
Pao H T, Tsai C M, 2011. Multivariate Granger causality between CO2 emissions, energy consumption, FDI (foreign direct investment) and GDP (gross domestic product): Evidence from a panel of BRIC (Brazil, Russian Federation, India, and China) countries. Energy, 36(1): 685-693.

DOI

[28]
Peng S, Ciais P, Maignan F, et al, 2017. Sensitivity of land use change emission estimates to historical land use and land cover mapping. Global Biogeochemical Cycles, 31(4): 626-643.

DOI

[29]
Samie A, Deng X, Jia S, et al, 2017. Scenario-based simulation on dynamics of land-use-land-cover change in Punjab Province, Pakistan. Sustainability, 9(8): 1285.

DOI

[30]
Shen L, Sun Y, 2016. Review on carbon emissions, energy consumption and low-carbon economy in China from a perspective of global climate change. Journal of Geographical Sciences, 26(7): 855-870.

DOI

[31]
Song M, Guo X, Wu K, et al, 2015. Driving effect analysis of energy-consumption carbon emissions in the Yangtze River Delta region. Journal of Cleaner Production, 103: 620-628.

DOI

[32]
Sun P, Xu Y, Yu Z, et al, 2016. Scenario simulation and landscape pattern dynamic changes of land use in the Poverty Belt around Beijing and Tianjin: A case study of Zhangjiakou city, Hebei Province. Journal of Geographical Sciences, 26(3): 272-296.

DOI

[33]
Tobler W R, 1979. Cellular geography. In: Philosophyin Geography. Dordrecht: Springer, 379-386.

[34]
Verburg P H, Soepboer W, Veldkamp A, et al., 2002. Modeling the spatial dynamics of regional land use: The CLUE-S model. Environmental Management, 30(3): 391-405.

PMID

[35]
Wang C, Wu K, Zhang X, et al, 2019. Features and drivers for energy-related carbon emissions in mega city: The case of Guangzhou, China based on an extended LMDI model. PloS One, 14(2): e0210430.

DOI

[36]
Wang M Q, Lu Y L, 2014. The study on eco-compensation in restricted development zones under the framework of the main functional area. Advanced Materials Research, 962: 2145-2148.

[37]
Wang Y, Song Q, He J, et al, 2015. Developing low-carbon cities through pilots. Climate Policy, 15(Suppl.1): S81-S103.

DOI

[38]
Xu F, Huang Q, Yue H, et al, 2020. Reexamining the relationship between urbanization and pollutant emissions in China based on the STIRPAT model. Journal of Environmental Management, 273: 111134.

DOI

[39]
Xu Q, Yang R, Dong Y X, et al, 2016. The influence of rapid urbanization and land use changes on terrestrial carbon sources/sinks in Guangzhou, China. Ecological Indicators, 70: 304-316.

DOI

[40]
Yang R, Pan Y, Xu Q, 2020. Space diversification process and evolution mechanism of typical village in the suburbs of Guangzhou: A case study of Beicun. Journal of Geographical Sciences, 30(7): 1155-1178.

DOI

[41]
York R, Rosa E A, Dietz T, 2003. STIRPAT, IPAT and ImPACT: Analytic tools for unpacking the driving forces of environmental impacts. Ecological Economics, 46(3): 351-365.

DOI

[42]
Zhang D, Fu M, Tao J, et al, 2013. Scenario simulation of land use change in mining city based on CLUE-S model. Transactions of the Chinese Society of Agricultural Engineering, 29(12): 246-256. (in Chinese)

[43]
Zhang S, Zhao T, 2019. Identifying major influencing factors of CO2 emissions in China: Regional disparities analysis based on STIRPAT model from 1996 to 2015. Atmospheric Environment, 207: 136-147.

DOI

[44]
Zhao C, Chen B, Hayat T, et al, 2014. Driving force analysis of water footprint change based on extended STIRPAT model: Evidence from the Chinese agricultural sector. Ecological Indicators, 47: 43-49.

DOI

Options
Outlines

/

〈 〉