Assessing Spatiotemporal Dynamics of Land Use and Cover Change and Carbon Storage in China’s Ecological Conservation Pilot Zone: A Case Study in Fujian Province

Abstract

Many strategies have been put forward to seek green and low-carbon development, some of which are achieved through land use and cover change (LUCC). A series of land management policies related to LUCC and corresponding changes in carbon dynamics were released with the implementation of the Ecological Conservation Pilot Zone Program (ECPZP) in China. We explored the spatiotemporal dynamics of LUCC and carbon storage in the first ECPZP implementation region (Fujian province) at the time before and after ECPZP implementation using a simplified carbon pools model and quantified the relative impacts of human activities and climate change on net primary productivity (NPP) employing residual analysis. This can fill the gap of land use and vegetation changes and the corresponding carbon dynamics in the ECPZP region and can serve as a reference for future land management policy revisions and ECPZP project extensions. The results showed that: (1) In 1990–2020, woodland, cultivated land, and grassland were the leading land use type in Fujian province. The area of LUCC was 11,707.75 km², and it was predominantly caused by the conversion from cultivated land to built-up land, and the interconversion between woodland and grassland. (2) An increase of 9.74 Tg in carbon storage was mainly caused by vegetation conversion from 1990 to 2020. (3) The statistically significant increased area of climate change-induced NPP was 2.3% primarily in the northwest, but the decreased area of it statistically significantly was only 0.1%. Correspondingly, the increased area of statistically significant human activity-induced NPP was 8.7% primarily in the southeast, but the decreased area of statistically significance was 6.5%, mostly in the central region. In addition, the statistically significant areas of NPP caused by the combination of human activities and climate change differed by 1.8%. To sum up, ECPZP makes full use of the vertical mountain landscape and property right reform to effectively secure ecological space and local income. Moreover, urbanization-related policies are an essential impetus for LUCC and carbon balance. The impact of other built-up land expansion on environmental change needs to be paid particular attention to. Moreover, land-use activities in the centre of the study region that are not conducive to NPP growth should be judiciously assessed in the future.

1. Introduction

The terrestrial ecosystem is a vast carbon pool, which plays an essential role in regulating the global carbon cycle and coping with the challenges of climate change [1,2,3]. The variations caused by carbon fluctuations of soil and vegetation in those carbon pools are profoundly affected by land use and cover change (LUCC) [4]. Carbon storage capacity is an indicator of the carbon pool size, which has no fixed value and varies significantly in different ecosystem types [5,6,7]. LUCC influences vegetation carbon storage directly by altering ecosystem type, and it regulates soil carbon storage progressively by affecting soil physicochemical, biological properties, and carbon input from vegetation litter and roots [8,9,10,11]. Over the past 200 years, the terrestrial ecosystem has suffered considerable losses in carbon storage with the industrial revolution and population growth [12]. LUCC, such as converting green land to urban areas and cropland, reduces carbon storage and messes up carbon flux exchange, causing pressure on climate mitigation [5,13]. Some studies simulating future land use scenarios reported that the continuous loss tendency of carbon storage would not be effectively reversed in the coming decades due to the expansion of urban areas and cropland [14,15,16,17]. Thus, digging into the effects of LUCC on carbon storage is crucial for global carbon management.

Numerous studies involving carbon storage related to LUCC have been conducted, which can be divided into two perspectives. One is to analyze the characteristics of LUCC, and the corresponding carbon storage dynamics from the historical perspective [18,19,20]. The other is to explore the effects on carbon storage from the future perspective by setting future LUCC scenarios, such as the future urbanization scenarios [21,22]. A growing number of studies results indicated that conducting proper land use planning and management measures [23,24], such as afforestation [25,26], designation of protected areas [27,28], agroforestry systems [29,30], pasture management [31], and ecological restoration [32], can effectively cope with the loss of carbon storage in terrestrial ecosystem induced by LUCC [33,34,35].

Methods employed to estimate the effects of LUCC on carbon storage were dominated by field surveys, remote sensing, and model simulations [35]. Field surveys can provide precise data, and are suitable for small-scale extent studies with limited budgets, but they are costly for large-scale extent [36]. Remote sensing is widely applied due to its low cost and broad coverage [37], and it is well suited for the investigation of ecosystem carbon storage when combined with field survey data [38]. Nevertheless, the accuracy of carbon storage assessment is affected by the spatial and temporal dimensions of remote sensing data [39]. Model simulations have grown in popularity as a means of geographically and explicitly assessing carbon storage [35,40] at the island scale [41], watershed scale [42], national scale [43], and global scale [44]. In particular, the InVEST (Integrated Valuation of Ecosystem Services and Tradeoffs) model stands out with a reliable and straightforward evaluation method [45,46]. The InVEST model quantifies the total carbon storage with a basic carbon cycle based on carbon density pools [47,48], including vegetation carbon, soil carbon, and dead organic carbon. Changes in those carbon pools induced by LUCC could be evaluated using the InVEST model [49].

The United Nations Framework Convention on Climate Change acknowledges that land use can contribute substantially to the mitigation of climate change, such as enhancing soil carbon sequestration and forest carbon storage. In China, the Ecological Conservation Pilot Zone Project (ECPZP) was practiced as a national plan to put systems in place for promoting ecological progress, and provide a replicable experience for other regions [50]. The land management contents of the ECPZP include but are not limited to ecological conservation redlines, cultivated land protection, territorial space planning, collective forestry tenure reform, and ecological restoration such as land reclamation and afforestation. People’s response to economic opportunity drives land use change under institutional regulation, and ECPZP provides an chance for the regulation of economic opportunity and institutional environment [51]. Exploring the LUCC and corresponding carbon storage dynamics induced by human activities related to land management in ECPZP is vital for the promotion and improvement of ECPZP as well as global carbon management. Currently, less attention has been focused, however, on ECPZP projects in China.

In this study, we selected Fujian province as the study region, and it is one of the earliest regions to implement ECPZP. The spatiotemporal dynamics of LUCC and carbon storage was evaluated during the implementation of ECPZP in Fujian province. We first analyzed LUCC from 1990 to 2020, and quantified carbon storage affected by LUCC. It is noteworthy that the dominant land use type in Fujian province is woodland, and vegetation change may be a major aspect of LUCC. Consequently, we further explored the relative impact of land management-related human activities on NPP to measure the vegetation carbon change in the ECPZP region. The research findings on the dynamical characteristics of LUCC and related carbon storage in ECPZP can be reference for the adjustment of future land use management. Moreover, the land management measures in ECPZP and these findings can also provide references in carbon management for other regions with similar natural and social conditions in China or around the world.

2. Materials and Methods

2.1. Study Region

Fujian province (23°33′–28°20′N, 115°50′–120°40′E) is located on the southeast coast of mainland China, bordering Zhejiang, Guangdong, and Jiangxi provinces, covering an area of 124,000 km² (Figure 1). The region is dominated by woodland, cultivated land, and grassland, and these land use types occupy more than 90% of the total area. Moreover, more than 48% of the area is evergreen broad-leaf forest, and about 15% is cultivated land and grassland, respectively. The overall topography of Fujian province is high in the northwest and low in the southeast. There are four major plains on the eastern coast, including Fuzhou Plain, Xinghua Plain, Quanzhou Plain and Zhangzhou Plain.

In 2001, Fujian province was one of the earliest provinces in China to put forward the Ecological Province Strategy aimed at ecological civilization. The Ecological Province Strategy can be considered as the predecessor of the ECPZP at the provincial level. In 2014, Fujian province was identified by the Chinese government as the first Demonstration Pilot Zone of Ecological Civilization in China. In 2016, Fujian province was listed as one of national ECPZP, therefore, the implementation of ECPZP in Fujian province can be divided into three periods, namely, the gestation period (before 2001), the proposal and development period (2001–2015), and the deepening period (after 2015). Under the guidance of the ecological civilization concept, both positive economic growth and ecological protection developments have been achieved in Fujian province. The GDP in Fujian province was 4.88 trillion yuan in 2021, and ecological achievements were widely lauded, such as the Changting county’s soil erosion prevention project and the construction of the Wuyishan National Nature Reserve.

2.2. Data Preparation

The primary supporting data of this study incorporated land use datasets, meteorological (precipitation and air temperature) datasets, and NPP datasets. Land use datasets (30 × 30 m) in 1990, 2000, 2005, 2010, 2015, and 2020 were retrieved from the Chinese Academy of Sciences Geography Science and Resource Institute (http://www.resdc.cn, accessed on 21 May 2022) [52]. The classification accuracy of the initial land use datasets was more than 90% [53]. According to the classification system of land use remote sensing monitoring data and the classification standards of the third national land survey in China, the land use datasets were reclassified into nine types: (1) Cultivated land: paddy field and dry field; (2) Woodland: evergreen, deciduous and mixed woodland; (3) Grassland: high coverage grassland, medium coverage grassland, and low coverage grassland; (4) Surface water: rivers and canals, reservoirs and ponds, and lakes; (5) Wetland: mudflats, marshes, and flood plain; (6) Urban built-up land: built-up land of cities and counties; (7) Rural residential land: rural settlements outside urban built-up land; (8) Other built-up land: such as transportation land, quarry, and saltern; (9) Bare land: land with sparse vegetation cover.

The monthly precipitation and air temperature datasets from 2000 to 2020 were obtained from the national Tibetan plateau data center (http://data.tpdc.ac.cn/en/, accessed on 21 May 2022) [54]. The raw datasets were projected and aggregated to form annual meteorological datasets with a spatial resolution of 1 km. As a variable reflecting carbon dynamic, we acquired and masked NPP data from annual MOD17A3H with a spatial resolution of 500 m over the last 21 years, and the accuracy of this dataset has been verified [5]. To ensure consistency with the spatial resolution of the meteorological data, the NPP datasets were resampled to a spatial resolution of 1 km.

2.3. Methodology

2.3.1. Simulation of Carbon Storage

The InVEST model was applied to simulate carbon storage by five carbon pools in each land use type: (1) aboveground vegetation carbon storage (branches, trunks, leaves, bark, and other living plant materials), (2) belowground vegetation carbon stock (living root system), (3) dead carbon storage (dead matter and litter), (4) soil organic carbon storage, and (5) unreleased carbon from timber cutting [22,55]. Generally, the fifth carbon pool is less available than the other four. Thus, the fifth carbon pool is rarely included in the carbon storage calculations in published papers, nor in this study.

Carbon density data for each land use type was collected from published literature, focusing on the study areas with similar climatic and geomorphological conditions in Fujian province [41]. Zero values were used to avoid misestimation of carbon storage for the categories in which no valid value was collected. Considering that built-up land has a certain vegetation cover, we multiplied the proportion of built-up land that was covered by vegetation by the average vegetation carbon density of woodland and grassland vegetation types to calculate the carbon density of built-up areas. Given that proportional vegetation cover in built-up areas changed from year to year, this calculation was repeated for each of the six time periods included in our study to determine the carbon density of built-up areas in each time period. Consequently, the soil organic carbon density of built-up land in this paper was determined to be one third of that of woodland according to [56], as shown in

Table 1. Carbon storage per unit area of each land use type (kg/m²).

Land Use Type	AGC	BGC	SOC	DC	Total	References
Cultivated land	0.65	0.43	2.98	0.1	4.16	[57,58,59]
Woodland	6.59	1.22	4.3	0.59	12.7	[58,59,60]
Grassland	0.07	0.31	3.22	0.02	3.62	[58,59,60,61]
Surface water	0	0	2.51	0	2.51	[58]
Wetland	0.57	0.55	4.5	0.15	5.77	[57,58,60]
Urban built-up land (1990)	0.15	0.03	1.42	0	1.6	[56,58,62]
Urban built-up land (2000)	1.1	0.25	1.42	0	2.77
Urban built-up land (2005)	1.2	0.28	1.42	0	2.9
Urban built-up land (2010)	1.37	0.31	1.42	0	3.1
Urban built-up land (2015)	1.43	0.33	1.42	0	3.18
Urban built-up land (2020)	1.49	0.34	1.42	0	3.25
Rural residential land	0	0	1.42	0	1.42
Other built-up land	0	0	1.42	0	1.42
Bare land	0	0	0	0	0

. Based on the above, the carbon density and total storage calculation method for land use type $i$ is expressed as Equations (1) and (2).

$$C_i=C_{i_{above}}+C_{i_{below}}+C_{i_{dead}}+C_{i_{soil}}$$

(1)

$$C_{total}={\displaystyle \sum }_{i=1}^N{C}_i\times S_i$$

(2)

where $i$ is a land use type; $C_i$ is the carbon density of land use type $i$ (unit: kg/m²); $C_{i_{above}}$, $C_{i_{below}}$, $C_{i_{dead}}$ and $C_{i_{soil}}$ are the aboveground carbon density (AGC), the belowground carbon (BGC) density, the dead carbon (DC) density and topsoil organic carbon (SOC) density of land use type $i$, respectively; $C_{total}$ is the total carbon storage (unit: kg) of the entire study area; $S_i$ is the area of land use type $i$ (unit: m²); and $N$ is the number of land use types, which is equal to 9 in this paper.

2.3.2. Calculation of Potential NPP and Human Activities-Affected NPP

The Potential NPP (PNPP) reflects the impact of climate change on NPP [63]. Thornthwaite memorial model has been widely used as one of the PNPP calculations [64], the formula is as Equations (3) and (4). The actual NPP (ANPP) is the NPP influenced by climate-driven and non-climate factors, which was collected from MOD17A3H; and the impact of human activities on NPP (HNPP) can be computed by subtracting the PNPP from the ANPP of the corresponding year. It should be noted that the PNPP in this study predicted an ideal condition determined only by climatic conditions, and changes in HNPP include all non-climatic factors [65].

$$PNPP=3000\left(1-e^{-0.0009695\left(E-20\right)}\right)$$

(3)

$$E=1.05p/\sqrt{1+{(1+1.05p/\left(3000+25t+0.05t^3\right))}^2}$$

(4)

where $E$ is the annual actual evapotranspiration (unit: mm), $t$ is the annual average temperature (unit: °C), $p$ is the annual precipitation (unit: mm), and PNPP is annual PNPP (unit: g/m²/year).

2.3.3. Analyzation of the Trend in NPP

Theil Sen’s slope has received enough attention as a statistical method to analyze linear trend of ecological variables [66]. Combined with the Mann Kendall nonparametric test, the statistical significance of the variable trend change could be calculated [67]. We used this method to analyze the trend of ANPP, HNPP, and PNPP pixel by pixel. Furthermore, both Theil Sen’s slope and Mann–Kendall test was measured using R software [68]. The Theil Sen’s slope calculation is presented as follows.

$$slope=\frac{n\times {{\displaystyle \sum }}_{i=1}^{n}i\times NP{P}_i-{{\displaystyle \sum }}_{i=1}^{n}i {{\displaystyle \sum }}_{i=1}^{n}NP{P}_i}{n\times {{\displaystyle \sum }}_{i=1}^n{i}^2-{\left({{\displaystyle \sum }}_{i=1}^{n}i\right)}^2}$$

(5)

where $n$ represents timescales from 2000 to 2020, the $NP{P}_i$ refers to ANPP, HNPP, and PNPP values in year $i$, respectively. Here, $slope>0$ indicates ANPP, HNPP, and PNPP growth, respectively, and vice versa.

3. Results

3.1. LUCC from 1990 to 2020

From 1990 to 2020, Fujian province was dominated by woodland, cultivated land, and grassland (Figure 2). The woodland area occupied over 61% of the total area. As the second-largest land use type, cultivated land occupied more than 17%, and which was scattered in the northwest (such as Nanping city) and the southeastern coastal plain region. Additionally, the grassland area exceeded 15%, mainly dispersed in the northeast (such as Zhenghe county and Zherong county), the central part (such as Youxi county), and the south (such as Anxi county and Pinghe county). Surface water, wetland, and bare land were mainly distributed in rivers and offshore, which accounted for about 1%, 0.3%, and 0.08%, respectively. Built-up land area occupied about 2% and was concentrated in the southeast coastal plain and central urban regions of inland.

From 1990 to 2020, the area of cultivated land and grassland continued to decline, declining by 1.65% and 2.1%, respectively. The area of woodland first increased and then dropped, as a result, the proportion of woodland increased slightly (0.61%) within 30 years. The area of built-up land increased significantly from 1.87% to 4.50%. Over the past 30 years, cultivated land, woodland, and grassland have been the sources of growth area of built-up land, and these changes mainly occurred in the coastal plain and inland mountainous areas.

3.1.1. Spatiotemporal Pattern of LUCC in Gestation Stage of ECPZP

Compared with 1990 (Figure 2), the proportion of woodland increased to 62.62%, built-up land increased to 2.03%, and cultivated land decreased to 18.34% in 2000. The most significant change in land use type was grassland in 1990–2000, with the ratio falling to 15.64%. In addition, the proportion of surface water and unused land remained stable.

A total area of 5483.91 km² changed from 1990 to 2000 in Fujian province (Figure 3A). The conversion area from grassland to woodland accounted for 58.10%, and the increased woodland was mainly distributed in the center and north. Conversely, the conversion area from woodland to grassland accounted for 24.72%, and the decreased grassland was concentrated in Gutian county, Minqing county, Minhou county, and Jin’an district. The conversions were scattered, including cultivated land to woodland, woodland to cultivated land, and grassland to cultivated land, accounting for 3.79%, 3.38%, and 1.34%, respectively. The conversion from cultivated land to built-up land accounted for 3.21% and was concentrated in the coastal region.

It was evident that the mutual conversion of woodland and grassland, characterized by the transition from grassland to woodland, was the dominant mode of land use change at this stage. In addition, the reduction of dry field at this stage contributed significantly to the loss of cultivated land. The area of urban built-up land, rural residential land and other built-up land increased slightly.

3.1.2. Spatiotemporal Pattern of LUCC in Proposal and Development Stage of ECPZP

The duration in proposal and development stage of ECPZP was approximately 15 years, and about 4220.64 km² of LUCC occurred during this stage (Figure 3B–D). The conversion area of each land use type tended to decline in 2000–2005, 2005–2010, and 2010–2015, which accounted for 2665.49 km², 1168.23 km², and 386.92 km², respectively. Most of the increased built-up land was viewed in the delta zone of Quanzhou city, Xiamen city and Zhangzhou city. The built-up land was gradually expanding southward along with the coastal cities. The increased surface water was mainly presented in the southern coastal regions. The decreased sites in cultivated land were relatively concentrated in the urban areas of Fuzhou city and Quanzhou city.

During this stage, cultivated land, grassland, woodland, and wetland decreased by 1.05%, 0.53%, 0.16%, and 0.04%, respectively (Figure 2). The reduction area of paddy field was the main part for losing cultivated land in acreage. In addition, built-up land and surface water expanded by 1.69% (including 1.04% for other built-up land and 0.51% for urban built-up land) and 0.09%, respectively.

Frequent mutual conversion occurred among different land use types in Fujian province, but these conversions exhibited variability at different times of this stage. During this stage, most of the changes in cultivated land, grassland, surface water, urban built-up land and other built-up land took place between 2000 and 2005. There was a decrease of 0.69%, 0.09% and 0.52% in cultivated land, woodland and grassland between 2000 and 2005 (Figure 2). In 2000–2005, the surface water area expended by 0.33%, and the built-up land expanded by 0.97%, including 0.47% increase in urban built-up land and 0.5% increase in other built-up land. Additionally, cultivated land decreased by 0.15% in 2005–2010, and by 0.13% in 2010–2015. Moreover, surface water decreased sharply (0.23%) in 2005–2010. Other built-up land increased by 0.52% in 2005–2010, and rural residential land decreased by 0.14% in 2005–2010.

The area of LUCC gradually decreased during the proposal and development stage of the ECPZP. The conversions of grassland to woodland, and woodland to other built-up land accounted for a significant proportion of the changed area between 2000 and 2010. The loss of cultivated land, especially paddy field, was dominated by the conversion of cultivated land to built-up land. The conversions of surface water to other built-up land, and woodland to grassland were as critical changed land use types between 2005 and 2010.

From 2000 to 2015, the rapid growth of built-up land mainly originated from the encroachment on cultivated land and woodland, which distributed in Fuzhou city, Quanzhou city, Xiamen city, Zhangzhou city, and Longyan city. Additionally, from 2000 to 2010, the interconversion between built-up land and surface water was observed in the coastal mudflats of Xiang’an district and Zhangpu county and Dongshan county.

3.1.3. Spatiotemporal Pattern of LUCC in Deepening Stage of ECPZP

In 2020, the surface water, wetland, and built-up land increased by 0.24%, 0.21%, and 0.78%, respectively, compared with them in 2015 (Figure 2). However, the proportion of cultivated land decreased by 0.48%, and woodland decreased by 0.74% in 2015–2020. The delta regions of Xiamen city, Quanzhou city, Zhangzhou city, Nanping city, and the urban regions of Fuzhou were the regions where the area of built-up land has increased significantly (Figure 2). The increased surface water was distributed in the coastal mudflats of Dongshan county, Zhangpu county, Xiang’an district, and Fuqing city. Otherwise, the newly constructed inland reservoirs also increased the proportion of surface water in the total changing area.

The total conversion area was 2003.20 km² from 2015 to 2020 (Figure 3E). The conversion area from cultivated land to other built-up land ranked first (16.46%), and the next two were the conversion from woodland to grassland (10.24%), and the conversion from woodland to built-up land (12.58%). In addition to the changes mentioned above, the conversion of other built-up land to surface water also contributed 6.06%. The sources of growing built-up land, especially new other built-up land, were still mainly cultivated land and woodland at the deepening stage of ECPZP.

3.2. Spatiotemporal Dynamics of Carbon Storage from 1990 to 2020

3.2.1. Dynamics of Carbon Storage

The spatial distribution pattern of carbon density in Fujian province was roughly similar from 1990 to 2020. Most regions were covered by high carbon densities, depicting a median-high-low distribution from northwest, centre, to southeast. The low-value regions gradually expanded in a contiguous form in the coastal plain, especially the surrounding regions of the built-up land where conversion occurred more intensively (Figure 4). The expansion of the low-value regions was relatively fragmented in the mountainous inland, and the expansion densely distributed in central urban built-up land.

According to

Table 2. Carbon storage of land use type (Tg).

Land Use Type	1990	2000	2005	2010	2015	2020
Woodland	945.04	968.54	967.56	965.98	965.95	961.93
Grassland	75.80	68.95	66.69	66.63	66.60	67.03
Surface water	2.76	2.80	3.83	3.10	3.08	3.86
Wetland	2.72	2.61	2.49	2.35	2.35	3.81
Bare land	0.00	0.00	0.00	0.00	0.00	0.00
Cultivated land	93.52	92.94	89.49	88.37	87.59	85.81
Urban built-up land	0.91	1.82	3.60	3.92	4.05	4.67
Rural residential land	1.70	1.80	1.79	1.81	2.04	2.23
Other built-up land	0.73	0.77	1.64	2.54	2.59	3.58
Total	1123.17	1140.23	1137.09	1134.70	1134.24	1132.91

, the total carbon storage in 30 years increased by 9.74 Tg, but the increase rose rapidly in the first decade and then continued to rise slowly (Figure 5A). Over 30 years, vegetation (aboveground and belowground) contributed 10.20 Tg to carbon storage growth, while soil weakened 1.04 Tg of carbon storage. The variation of vegetation carbon storage was mainly caused by the change of woodland area, while the variation of soil carbon storage was mainly dominated by the change area of woodland and grassland (Figure 5B,C). The carbon storage changed significantly in 1990–2005 and 2015–2020 (Figure 5).

In the study period, the carbon storage decreased continuously only in cultivated land, which decreased by 7.71 Tg. The carbon storage of built-up land continued to rise and was enhanced by 7.14 Tg. In addition, carbon storage in other land use types changed at different times. Carbon storage in grassland and wetland gradually decreased from 1990 to 2015 and increased from 2015 to 2020. As a result, carbon storage in grassland decreased by 8.77 Tg, and increased by 1.09 Tg in wetland. Carbon storage in surface water fluctuated over 30 years with an increase of 1.10 Tg. At any given time, the largest share of carbon storage was in woodland (no less than 84%), followed by cultivated land and grassland. Moreover, the carbon storage of other types of land was relatively small (Table 2).

3.2.2. Effects of Land Use Transition on Carbon Storage

In the gestation stage of ECPZP (1990–2000) (Figure 6A), the conversion from grassland to woodland was a primary driver of increased carbon storage (28,930.26 Gg). In addition, the conversion from cultivated land to grassland and the enhanced vegetation coverage in urban built-up land also contributed to the carbon storage growth with a total of 1776.15 Gg and 662.57 Gg. Conversely, the conversion from woodland to grassland and cultivated land resulted in losses of carbon storage, which accounted for 12,306.89 Gg and 1583.08 Gg, respectively.

In the proposal and development stage of ECPZP (2000–2015) (Figure 6B), the carbon storage loss was 5.99 Tg, due to the conversion from cultivated land and woodland to built-up land with losses of 2602.45 Gg and 6587.41 Gg, respectively. In particular, the carbon storage losses induced by the expansion of other built-up land to cultivated land and woodland reached 1654.32 Gg and 5380.67 Tg, respectively. However, the conversion from grassland and cultivated land to woodland increased carbon storage by 5745.99 Gg, and a new carbon storage (524.21 Gg) was formed owing to the increase of green coverage in urban built-up land.

In the deepening stage of ECPZP (2015–2020) (Figure 6C), the loss of carbon storage was mainly caused by the loss of woodland area, followed by the decline of cultivated land area. The conversion from woodland to built-up land reduced carbon storage by 3488.06 Gg, of which 2842.24 Gg carbon storage was lost due to the occupation of woodland by other built-up land. The degradation of woodland to grassland resulted in carbon emissions of 1862.77 Gg. In addition, the conversion from cultivated land to built-up land resulted in a loss of 1244.07 Gg of carbon storage, with the loss caused by the expansion of other built-up land accounting for 72.62% of the total. Nevertheless, returning farmland to forests, restoring grassland to forests, and restoring built-up land to forests, respectively, increased carbon storage by 155.74 Gg, 403.93 Gg, and 536.57 Gg.

In summary, the dynamics area of the woodland served as a barometer for changes in carbon storage. The expansion of urban built-up land may lead to increased carbon storage, but whether the urban vegetation coverage has been planned was the premise.

3.3. NPP Changes in Fujian Province

According to our results, vegetation was the main driving force of LUCC and carbon change. However, the effect of human activities on vegetation carbon changes is still unclear, which is not conducive to the continued implementation and revision of future land management policies. Therefore, we next discussed the impact of human activities on changes in vegetation carbon.

3.3.1. Spatiotemporal Changes in ANPP

From 2000 to 2020, the mean ANPP in the study area varied from 0.78 kg C/m² to 0.91 kg C/m² with a mean value of 0.85 kg C/m², and the overall trend showed a slight but statistically insignificant decrease (Figure 7A). There was no significant change in the mean ANPP time series by the Mann-Kendall test method. Therefore, since the implementation of ECPZP in the study region, the ANPP value has been stable. Synthesizing the 21-year ANPP value image employed mean method and classifying it into four categories based on natural breakpoints, the smaller ANPP value to larger one showed a stepwise distribution from northwest to southeast. And there was apparent spatial heterogeneity (Figure 7B). Combining the Mann–Kendall test (95% significance) with Theil Sen’s slope results, we found that the growth area of ANPP accounted for 40.85% of total area. The significant growth area of ANPP value accounted for 16.04% of the total area, which concentrated in Anxi county, Yongchun county, Zhaoan county, Pinghe county, and Yunxiao county in the southeast. A slight increase in the area of ANPP occupied 24.81% with a scattered distribution. In contrast, the decrease in ANPP accounted for 59.13% of the total area. The significant decline of ANPP occupied 10.01% and was mainly spread in Fuzhou city, Dehua county, and Youxi county. A slight decline of ANPP accounted for 49.12%, which was widely distributed throughout the study region (Figure 7C).

3.3.2. Relative Effects of Climate Change and Human Activities on NPP

Regrading to the method mentioned in [69], the relative contributions of human activities and climate change to NPP changes were explored. The area of human activities-induced NPP decline accounted for the most proportion with about 42.2%, the area of human activities-induced NPP growth accounted for 16.5%, and the area of climate change-induced NPP growth accounted for 15.1%, respectively. Additionally, there was little difference in the proportion between the increase (9.0%) and decrease (9.3%) area of NPP caused by climate change and human activities. The area of NPP decreasing caused by climate change was the least area at 6.6% (Figure 8A). The spatial distribution of different dominant effects in NPP was hierarchical. The NPP growth areas dominated by climate change were mainly distributed in the northwest, while both the NPP decrease area and increase area dominated by human activities were mainly distributed in the centre and southeast.

Combined with the Mann–Kendall significance test results (Figure 8B), human activities and climate change significantly affected the increase and decrease of NPP value, and the area of increase was 1.8% larger than the area of decrease. From a statistical point of view, human activities and climate change jointly dominated the changes in vegetation carbon in the study area, and the effects of human activities on NPP changes were more evident than climate change. Climate change significantly influenced the increase in the NPP value in the northwest, while human activities significantly affected the increase in the NPP value in the southeast. However, human activities did not significantly cause a large decline in the NPP value in the central region.

4. Discussion

4.1. Policy Drivers and Implications of Land Use and Managements

4.1.1. Policy Drivers of Land Use and Managements

Different policies in different periods of ECPZP (Figure 9) can shape various characteristics of LUCC on socialist public ownership of land. From 1990 to 2000, the conversion from grassland to woodland was primarily due to the construction of a comprehensive coastal protection system implemented in 1988, and the 357 afforestation and greening project implemented in 1989. These actions made grassland and barren hills restored to woodland and shrubs. In the 1990s, Quanzhou’s Southeast Core Regional Development Plan began granting more flexible economic management authority to Xiamen, which facilitated the conversion of coastal plains from cultivated land to built-up land. Since China’s accession to the WTO in 2000, the excellent port conditions have enabled the development of the coastal foreign trade economy in Fujian province. In addition, the Special Plan for Urban Development and Urbanization of Tenth Five-Year Plan in Fujian Province was published in 2002, and the Taiwan Strait West Coast Economic Zone in 2004 was established, which were further promoting expansion of coastal cities. The implementation of Small-Town Strategy in 2010, the establishment of China (Fujian) Pilot Free Trade Zone in 2014, and the Fujian New Urbanization Plan (2014–2020) in 2014 were all promoted the economic development of the coastal plain and inland mountainous areas, and this stimulated the expansion of urban regions. Fortunately, ongoing tree-planting projects, strict cultivated land occupation and compensation balance policies, high-standard territorial space planning, construction of the Ecological Civilization Pilot Zone, and forest tenure reforms have been helpful for ecological protection in Fujian province. In summary, there are three driving factors which contribute significantly to land use change in Fujian province. First, the implementation of specific land use policies tends to directly change the type of land use under the basis of public ownership of land in China, such as afforestation and restriction of non-agriculturalization of cultivated land. Second, socio-economic development plans, such as urban development plans, guide the market to refine the division of labor and create space for people to pursue economic opportunities, so as to create new land uses under market and policy regulation [51]. Finally, Fujian province has numerous ports and product processing enterprises, and as an important node of the global economic network, globalization has strengthened economic opportunities in Fujian province.

4.1.2. Implications from ECPZP

The work of ECPZP in Fujian province was fruitful in balancing the depletion of cultivated land, ecological protection and economic growth. Combined with publicly available information, we summarized some implications in land management that may be constructive for the sustainable development of other places (Figure 9).

(1) In mountainous regions with scarce cultivated land and abundant woodland, stereoscopic agriculture is conducive to achieving a win-win situation in terms of economic and ecological benefits by making full use of mountain resources. For example, ecological public welfare forests and water-conserving forests were planted on the mountain tops, and fruits, grass, and edible mushrooms were planted on the mountainsides, such as in Ningde city. Planting cash crops and raising poultry in the forest understory are also manifestations of stereoscopic agriculture, such as in Zhangzhou City. These measures can prevent forest land from being further cleared and both economic income and carbon storage are increased.

(2) One of the core contents of ECPZP is to transform ecological resources into economic resources. That is, lucid waters and lush mountains are invaluable assets. The performance evaluation criteria of local officials have changed from being based solely on GDP to including more green factors in the performance sector [70]. Additionally, the unsustainability of rapid economic development at the cost of enormous energy consumption and environmental pollution has prompted the need for local governments to attract environmentally friendly industries for local economic development [70]. The property rights of forest land resources can be reasonably allocated to villagers by forest tenure reform. In Wuping county (the first county to reform forest tenure), the clear property rights have positively promoted the enthusiasm of residents to manage forestry and their family income. More importantly, the reform has significantly increased the income of proprietors from forest products, the forest economy, and transfer income (e.g., ecological compensation for ecological forests) [71].

(3) It is necessary to ensure the forest land area and protect the vegetation coverage from reducing, so that the function of carbon storage in woodland and grassland can be better played. The depletion of cultivated land and natural land (woodland, grassland, surface water) by other built-up land expansion is noteworthy. In the early stages of economic development, the process of urbanization was synchronized with industrialization, which led to the loss of cultivated land, woodland and grassland. Sometimes, the depletion of natural land and semi-natural land (cultivated land) caused by the expansion of other built-up land was more prominent than that caused by the expansion of urban built-up land and rural residential land. The damage to cultivated land and natural space induced by the expansion of other built-up land, such as mining and road construction, also needs to be taken seriously.

4.2. Limitations and Future Research

The aboveground carbon density, belowground carbon density, dead carbon density, and soil organic carbon density of cultivated land, surface water, built-up land, and unused land in this paper were gathered concerning relevant literature. Existing studies have reported that the carbon density of vegetation varies with tree age [72], and these transitions between different tree species also bring about transitions in carbon storage, such as the conversion of broadleaf forests to plantations [73]. In this paper, these data may be collected and tested from a single time point, reflecting carbon density at some level of equilibrium, and may be limited in presenting the dynamics of carbon density in real time. However, some studies have shown that even if ignoring carbon density changes, the effect of LUCC on carbon storage dynamics can be well assessed [20,55,74,75]. In addition, the PNPP was estimated in this paper based on temperature and precipitation models. In the future, more variables, such as solar radiation and maximum temperature, can be included in the PNPP calculation framework. The HNPP in this study measures the NPP influenced by all non-climatic factors, which corresponds to the purely climate-driven NPP. Simplified HNPP calculations may not allow for the detailed analysis of the effect of specific human activities on NPP, such as abandonment, plant disaster prevention, and land management. Moreover, the scope of human activities involved in the Ecological Conservation Pilot Zone Program (ECPZP) is also relatively wide. Accordingly, we explored the impact of ECPZP on carbon dynamics by excluding the pure climate impact. In the future, human activities can be subdivided to explore the findings on HNPP in more detail, which may provide more guidance for specific policy decisions.

5. Conclusions

This study provided spatiotemporal dynamics of LUCC and carbon storage, and the effects of human activities on NPP in Fujian province. We analyzed the spatiotemporal dynamics of LUCC and carbon storage at different periods of ECPZP and discussed the relative effects of human activities and climate change on NPP for future land management policies decision. The main conclusions are as follows.

(1) The control of ecological land areas, such as woodland and grassland, has witnessed the efficient management of land use in the three stages of ECPZP. From 1990 to 2020, woodland and built-up land were the main types of increased land use in Fujian province. The conversions from cultivated land, woodland and grassland to built-up land were the main types of LUCC, which occurred in coastal plains and inland urban areas in the past 30 years. Land use policies are driving forces to be reckoned with in promoting LUCC. The total area of LUCC increased significantly from 2015 to 2020, especially the increase of other built-up land (39.59% of the total increase). The increase of other built-up land at this stage was corroborated by the 2014 New Urbanization Plan of Fujian province, which focused on promoting the construction of transportation conditions in small and medium-sized cities.

(2) The carbon storage in Fujian province increased slightly during the study period, and the vegetation change has made a significant contribution to the dynamic changes of carbon storage. Moreover, the expansion of other built-up land was the primary driving force for the loss of carbon storage. The disturbance effect of other built-up land expansion on carbon storage is supposed to be paid more attention to in the future. In addition, the negative impact of urban expansion on carbon storage can be partially offset by positive measures, such as increasing urban green coverage.

(3) Since the implementation of the ECPZP policy in 2000, the overall trend of NPP has remained stable, and human activities and climate change have jointly dominated the changes in NPP. Human activities had a greater impact on NPP, which was particularly evident in the southeastern part of the study region. In the central part of the study region, the decrease in NPP caused by human activities was quite large. Future land use policies ought to receive more attention in the central region.

(4) Making full use of a vertical mountain landscape is one of the efficient ways to solve the land use space in mountainous areas with poor cultivated land, and it is also a way to ensure that the carbon storage in the forest land does not decrease. In addition, converting ecological resources to economic resources is an effective way to realize the coordinated development of social-natural ecosystems.