Spatiotemporal Analysis of Carbon Emissions and Uptake Changes from Land-Use in the Yangtze River Delta Region ^†

Abstract

Land use change and energy consumption caused by human activities is the primary source of carbon emissions and a driver of climate change. The study focused on the Yangtze River Delta (YRD), using the China Land Cover Dataset (CLCD) to calculate the region’s carbon emissions from 1990 to 2020. Based on the Natural Segment Method, the spatial distribution of carbon emissions in the YRD region were constructed by dividing them into three categories: heavy, medium, and light. The results indicate that: (1) Carbon emissions of the YRD region was 594.02 million tons at the end of 2020, an increase of 468.53 million tons compared with that of 1990. The impervious surface was the major source of carbon emissions, accounting for more than 98.51% of the total, and woodland was the most important carbon sink, accounting for more than 91.32% of the total carbon uptake. (2) The carbon emissions increase rate over the 30-year period has risen from rapid to gradual, with the fastest rate of increase occurring between 2000 and 2010. (3) Differences in economic development and land type lead to spatial variability in carbon emissions. Regions with substantial emissions were predominantly located in coastal areas, indicating a trend toward shifting inland. The assessment of carbon emissions is helpful for designing emissions mitigation policies.

1. Introduction

With the proposal of the carbon neutrality goal, people have realized that global warming, primarily driven by carbon emissions, is one of the major environmental challenges to human social development [1].

In recent years, much research on carbon emissions and influencing factors at various scales has been conducted. On a national level, a comprehensive analysis of carbon sources and sinks in China’s terrestrial experimental ecosystems was conducted through model simulation and other methods [2], which were employed to study the size of carbon sources/sinks and the dynamic changes on the time scale. On the provincial level, land use and statistical data were combined to explore the differences in land use carbon emissions [3]. The characteristics of land use carbon emissions and spatial differences in the counties of Fujian were analyzed to propose strategies for low-carbon optimization of the national land space [4]. At the district and county scales, nighttime lighting data was used for estimating the carbon emissions of Chinese counties, analyzing their spatial and temporal variations and influencing factors [5]. In addition, some studies take watersheds or main functional areas as the research scope. The social network analysis (SNA) method classified land into six types to calculate land use carbon emissions and verify the regional carbon emissions spatial characteristics [6].

In this paper, we leveraged the China Land Cover Dataset (CLCD) from 1990 to 2020 to quantify the area of each land cover type. The Carbon Emission Factor (CEF) method was used to estimate carbon emissions, creating a spatial distribution map of carbon emissions in the Yangtze River Delta (YRD) region. The study aims to provide an assist for ecological construction and agricultural development.

2. Materials

2.1. Area of Study

As shown in Figure 1, theYRD is one of the most economically developed and urbanized regions in China, covering Shanghai, Jiangsu, Zhejiang, and Anhui. Characterized by its flat terrain and dense river network, the YRD boasts a highly developed economy. It is renowned for having the best development foundation, the most favorable institutional environment, and the strongest overall competitiveness in the country. The YRD holds a crucial strategic position in China’s modernization efforts, serving as a key driver of national development due to its advantageous geographic location and robust economic strength.

2.2. Land Cover and Energy Consumption Data

CLCD, which was created by 335,709 Landsat data from Google Earth Engine and applied Random Forest Classifier, was chosen for study. The dataset covers the yearly land cover information of China from 1985 to 2020 [7]. The China Energy Statistics Yearbook is an annual report published by the National Bureau of Statistics of China (NBSC) [8]. The yearbook is used to comprehensively record and present relevant data and changes in China’s energy sector. Data on energy consumption in the YRD region from 1990 to 2020 were selected.

3. Methods

3.1. Carbon Emissions Calculation Model

The CEF method is widely used for estimating carbon emissions, using the carbon content and emission factors of energy consumption or material transformation processes. The calculation of land use carbon emission is divided into direct and indirect carbon emission estimations. Direct carbon emissions are calculated as follows:

$$E_t=\sum e_t=\sum S_t\ast {\epsilon }_t$$

(1)

$E_t$is direct carbon emissions, $e_t$ is each land cover type carbon emissions, $S_t$ is the tract of land cover type, ${\epsilon }_t$ is the CEF for different land cover types.

Based on the results of the studies [9,10] and the specific characteristics of the study area, the CEF for each land cover type were identified (

Table 1. Carbon emission coefficients for land cover type. (kg/hm²·a).

Category	Carbon Emission Factor
Cropland	0.420
Woodland	−0.578
Shrubland	−0.578
Grassland	−0.021
Waterbody	−0.252
Bareland	−0.005
Impervious surfaces	0.000
Wetland	−0.252

). The formula for indirect carbon emission is as follows.

$$E_K=\sum E_k\ast {\phi }_k\ast {\gamma }_k$$

(2)

$E_K$ is the carbon emission of Impervious surfaces, $E_k$ is various energy sources consumption, ${\phi }_k$ is the factor of standard coal energy conversion, and ${\gamma }_k$ denotes the CEF of each energy. The energy conversion coefficients and carbon emission coefficients refer to existing studies as shown in

Table 2. Carbon emission coefficients for the energy source. (kgce/kg or kgce/m³).

Category	Standard Coal Conversion Factor	Carbon Emission Factor
Coal	0.713	0.756
Coke	0.971	0.855
Oil	1.428	0.586
Crude Oil	1.429	0.586
Gasoline	1.471	0.554
Kerosene	1.471	0.571
Diesel	1.457	0.592
Fuel Oil	1.429	0.619
Liquefied petroleum gas	1.714	0.504
Natural gas	1.214	0.448
Electricity	0.404	0.794

:

3.2. LMDI Factor Decomposition Method

The Logarithmic Mean Divisia Index (LMDI) factor decomposition method, which is widely used in the research of decomposition analysis of carbon emissions and influencing factors, can decompose multiple factors and eliminate residual terms. Economic development, population size, energy structure, and energy intensity are selected as indicators related to carbon emissions. The decomposition is conducted using the following formula:

$$C={\sum }_i C_i={\sum }_i F_i\ast S_i\ast G_i\ast P={\sum }_i {\displaystyle \frac{E_i}{E}}\ast {\displaystyle \frac{E}{Y}}\ast {\displaystyle \frac{Y}{P}}\ast P$$

(3)

$F_i$ is energy structure, indicating the proportion of i in energy consumption; $S_i$ is energy intensity, denoting energy consumption per unit of GDP; $G_i$ is economic development, indicating per capita GDP; P is population size; $C_i$ is carbon emissions; E is the consumption of total energy; $E_i$ is energy source consumption; and Y is GDP.

4. Results

4.1. Analysis of Carbon Emissions in Spatial and Temporal

The carbon emissions in the YRD region from 1990 to 2020 were analyzed by the CEF method. As shown in

Table 3. Carbon emissions of land cover in the YRD region (10² t).

	1990	1995	2000	2005	2010	2015	2020
Cropland	86,765.53	85,063.51	83,224.90	80,461.13	78,071.33	76,855.16	76,586.24
Woodland	−62,261.96	−62,945.80	−62,493.49	−63,074.18	−63,191.67	−61,437.02	−60,706.99
Shrubland	−9.58	−6.94	−7.21	−5.71	−4.30	−3.65	−3.46
Grassland	−8.81	−5.08	−3.88	−3.30	−3.46	−2.26	−1.19
Waterbody	−5885.02	−5712.55	−5998.92	−6355.48	−6371.17	−6305.04	−5874.06
Bareland	−1.38	−0.25	−0.10	−0.05	−0.03	−0.02	−0.02
Impervious surface	1,236,321.97	1,785,655.55	2,133,395.73	3,555,596.63	4,962,517.45	5,692,017.14	5,930,218.51
Wetland	−0.64	−0.14	−0.05	−0.02	0.00	0.00	0.00

, carbon emissions have risen in the YRD region over the last 30 years, escalating from 125.49 million tons in 1990 to 594.02 million tons in 2020. Impervious surface is the main contributor to carbon emissions, occupying an average of over 98% annually and exhibiting consistent year-on-year growth. In contrast, cropland, a minor source of emissions, displayed a declining trend. Regarding carbon sink, woodland is the main carbon sink, comprising more than 91% of the total carbon sink. The carbon sink of woodland, waterbody, and grassland demonstrated a slight downward trend, while wetland and bareland showed negligible carbon sink effects due to their limited areas. The carbon emissions growth rate experienced a rapid and then slow process, with the fastest growth between 2000 and 2010, peaking at more than 30%, and then gradually declining, reaching its lowest point in the final five years (Figure 2).

To compare and analyze the differences among cities, carbon emission is classified into three levels: heavy, medium, and light. As shown in Figure 3, the areas with high values are primarily in coastal cities. The regions exhibiting the lowest values are predominantly situated in the southwestern portion of the country, while the regions exhibiting medium values are primarily located in the central region. This suggests a tendency for areas with heavy carbon emissions to shift towards the interior.

4.2. Analysis of Carbon Emission Impact Factors

The LMDI factor decomposition method is used to calculate the effect value of each influencing factor. A positive effect value indicates that the factor drives carbon emissions, while a negative effect value suggests that the factor inhibits carbon emissions, as illustrated in Figure 4.

Between 1990 and 2020, there are obvious geographical differences in the direction and size of the driving and inhibiting factors affecting carbon emissions in the YRD region provinces. Economic development is the primary enabler for increased carbon emissions in each province. Conversely, energy intensity serves as the main suppressor, contributing to the reduction of carbon emissions. The driving effect of economic development on carbon emissions is stronger than the inhibiting effect of energy intensity. The impact of energy structure on carbon emissions varies across provinces, exhibiting both positive driving and negative inhibiting effects. The population size of each province has a positive driving effect on carbon emissions, with little fluctuation.

5. Conclusions

The reduction of carbon emissions is one of the most important ways of promoting ecological balance and sustainable regional development. In this paper, carbon emissions in the YRD region from 1990 to 2020 were calculated by combining land use data and the CEF method. The results indicate that impervious surface is the primary source of carbon emissions, occupying 98% of the total, while woodland as the main carbon sink occupies more than 91% of the total. The carbon emissions during the study period increased annually, from 125.49 million tons in 1990 to 594.02 million tons in 2020, with the highest growth rate of more than 30% during the period from 2000 to 2010. Subsequently, the carbon emissions growth rate gradually decreased, reaching its lowest point in the final five years of the study. It is noteworthy that the coastal cities have higher carbon emissions while the southwestern cities have lower carbon emissions. The reasons for the different spatial distributions of carbon emissions are analyzed. Recommendations for emission reduction in different regions are made to provide valuable insights for urban planning and environmental protection initiatives in the YRD region.