Which impacts more seriously on natural habitat loss and degradation? Cropland expansion or urban expansion?
Funding information: Fundamental Research Funds for the Central Universities, Grant/Award Number: 2662017PY063; National Natural Science Foundation of China, Grant/Award Numbers: 41371113, 41971240; Post-finance Project for National Social Sciences, Grant/Award Number: 19FGLB071; Post-finance Project for Philosophy and Social Sciences of the Ministry of Education, Grant/Award Number: 18JHQ081
Abstract
Natural habitat plays an important role in maintaining biodiversity. Both cropland expansion and urban expansion have an influence on natural habitat. However, it is not clear which one impacts more seriously on both the quantity and quality of natural habitat. This study compared the impacts of cropland expansion on the quantity and quality of natural habitat in China between 2000 and 2015 with the impacts of urban expansion. Map algebra in ArcGIS 10.6 was used to calculate the changes in the quantity of natural habitat, while the Integrated Valuation of Ecosystem Services and Trade-offs (InVEST) Habitat Quality model was used to assess the changes in its quality. The results indicated that cropland expansion led to a loss of 35,811 km2 of natural habitat, which was twelve-times as much as that from urban expansion. Furthermore, the area of the heaviest habitat degradation due to cropland expansion was 9,530 km2, which was eight-times as much as that due to urban expansion. Noticeably, the greatest impacts of cropland expansion on natural habitat mostly occurred in areas where the ecological environment is already vulnerable (namely, the resistance and resilience of ecosystems in response to external interference are weak), whereas the impacts of urban expansion were much less in these areas. This study highlights that the impacts of cropland expansion on both the natural habitat loss and degradation far exceeded the impacts of urban expansion. It is necessary to improve cropland protection policies to guarantee food security while ensuring little or no harm to natural habitat.
1 INTRODUCTION
Natural habitat plays an important role in maintaining biodiversity, climate regulation, and mitigating droughts and floods (Pfeifer et al., 2017; Yu , Hu, van Vliet, Verburg, & Wu, 2018). It is noting that the loss of natural habitat can directly lead to a decline in biodiversity (Watson & Venter, 2017). More than 32,000 species (including vertebrates, invertebrates, plants, fungi, and protists) are at risk of extinction and account for 27% of all species assessed in the 2020’s IUCN Red List (IUCN Red List, 2020). The number of threatened species in 2020 is three-times as much as were threatened in 2000 (IUCN Red List, 2020). Species richness has reduced by 13.6% and is projected to fall by a further 3.4% by 2100, under a business-as-usual scenario globally (Newbold et al., 2015). In addition, global ecosystem services are also at great risk because of extensive natural habitat conversion (Peters et al., 2019), such as carbon storage, soil retention, and sandstorm prevention (Mao et al., 2019; Molotoks et al., 2018; Tang, Ke, Zhou, Zheng, & Wang, 2020b). Thus, it is vital to identify the causes of natural habitat loss.
Land-use change is one of the important factors causing natural habitat loss (Cayuela, Lambrey, Vacher, & Miaud, 2015). Global land-use change led to a large loss of forest from 1982 to 2016: Brazil lost 8% of its forest an area of 385,000 km2, Argentina's forests were reduced by 25% with 113,000 km2 lost, and Paraguay's forests decreased by 34% with 79,000 km2 lost (X. P. Song et al., 2018). During this period, the global arid and semiarid drylands experienced large decreases in areas of vegetation, such as the southwestern United States, Inner Mongolia of China, and a large part of Australia (X. P. Song et al., 2018). This trend likely will worsen in the future: global land-use change is predicted to reduce natural habitat by 26–58% from 2005 to 2100 (Jantz et al., 2015). Furthermore, 39% of the terrestrial natural habitat is occupied by cropland and urban land, and another 37% is degraded and fragmented between 1700 to 2000 (Ellis, Klein Goldewijk, Siebert, Lightman, & Ramankutty, 2010). Alcamo et al. (2005) indicated that 10%–20% of natural grassland and forest will be occupied by agriculture and urban infrastructure by 2050. Therefore, exploring the impacts of land-use change on the natural habitat is increasingly important (C. He, Zhang, Huang, & Zhao, 2016; W. Song & Deng, 2017).
Previous studies have explored natural habitat loss caused by urban expansion during historic periods (Chao, 2009; C. He, Liu, Tian, & Ma, 2014) and in the future (Mao et al., 2018; Seto, Guneralp, & Hutyra, 2012). Seto et al. (2012) projected that global urban land cover would increase by 1.2 million km2 between 2000 and 2030, which would inevitably result in natural habitat loss. In addition, the ways to mitigate the trade-offs between economic development and natural habitat loss in rapid urbanization areas have also been explored. For example, Mcdonald, Güneralp, Huang, Seto, and You (2018) suggested that priority ecoregions where conservation investments could best mitigate biodiversity loss due to urban expansion should be established. C. W. Huang, Mcdonald, and Seto (2018) proposed that the conservation strategies, which facilitate public participation, as well as international aid and development, were useful for protecting natural habitat and biodiversity.
Cropland expansion, as a major land-use change, also has impacts on natural habitat (Lark, Salmon, & Gibbs, 2015; Tang, Ke, Zhou, Wang, & Koomen, 2020a; van Vliet, 2019). Global losses of natural areas of forest and shrubland are primarily attributed to cropland expansion (van Vliet, 2019). Lark et al. (2015) found that cropland expansion caused a great loss of grassland, with million ha lost (accounting for 77% of all new cropland) between 2000 and 2008 in the United States. Nine states in the Brazilian Legal Amazonia experienced large deforestation as a result of cropland expansion (Morton et al.,2006). Furthermore, the area of cropland expansion replacing natural habitat was predicted to be 3.5 × 108 ha in 2050, which would lead to eutrophication and natural habitat destruction (Tilman et al., 2001). Cropland expansion is also a noticeable issue in China. Specifically, since urban land replaces large areas of cropland, threatening food security (C. He, Liu, Xu, Ma, & Dou, 2017a; Kuang, Liu, Dong, Chi, & Zhang, 2016; van Vliet, Eitelberg, & Verburg, 2017), the Chinese Government has implemented a series of cropland protection policies (Y. Chen, Zhang, Guo, Lu, & Wang, 2018; Kong, 2014; Xin & Li, 2018). For example, the Requisition–Compensation Balance of Cropland Policy states that new cropland should be reclaimed to compensate for the lost cropland due to urban expansion (Liu et al., 2017a; W. Song & Pijanowski, 2014). Consequently, large areas of natural habitat, such as forest, grassland, and wetlands (IUCN, 2019), have been inevitably been converted into cropland in China (Ke et al., 2018; Ke & Tang, 2019; Zheng, Ke, Zhou, & Yang, 2019).
Both cropland expansion and urban expansion lead to natural habitat loss. However, it is not clear which one impacts more seriously on natural habitat loss in terms of its' quantity and quality. This study aims to explore the magnitude of impacts of cropland expansion on both the quantity and quality of natural habitat in China between 2000 and 2015 by comparing them with the impacts of urban expansion. Map algebra in ArcGIS 10.6 (W. Chen et al., 2020; C. WuChen, Huang, & Wei, 2020) was used to calculate the changes in the quantity of natural habitat due to cropland expansion, which was compared with the changes due to urban expansion. Furthermore, the Integrated Valuation of Ecosystem Services and Trade-offs (InVEST) Habitat Quality model (Sharp et al., 2016) was applied to assess the changes in the quality of natural habitat caused by cropland expansion, which was also compared with the changes caused by urban expansion.
2 MATERIALS AND METHODS
2.1 Data
2.1.1 Land-use/cover datasets
Land-use/cover datasets (LUCDs) of China in 2000 and 2015, with a resolution of 1,000 m, were obtained from the Resources and Environment Data Cloud Platform, Chinese Academy of Science (CAS) (2018). Land-use/cover types were classified into 25 categories in the LUCDs, which were produced by the visual interpretation from LANDSAT Thematic Mapper (TM)/Enhanced Thematic Mapper (ETM) images in the Albers projection (J. Liu, Liu, et al., 2005a; W. Song & Deng,2017).
The Habitats Classification Scheme, developed by the International Union for Conservation of Nature and Natural Resources (IUCN) (IUCN, 2019), was adopted for extracting natural habitat from the LUCDs. According to this scheme, the natural habitat datasets in China were generated by converting land-use/cover types in the LUCDs into the natural habitat types by the relationship shown in Table 1.
| Natural habitat typesa | Land-use/cover typesb |
|---|---|
| Forest | Forest land |
| Other forest | |
| Savanna | Open forest savanna |
| Shrubland | Shrub wood |
| Grassland | Dense grass |
| Moderate grass | |
| Sparse grass | |
| Wetlands | Stream and rivers |
| Lakes | |
| Reservoir and ponds | |
| Beach and shore | |
| Bottomland | |
| Swampland | |
| Saline | |
| Rocky areas | Bare rock |
| Desert | Sandy land |
| Gobi | |
Bare soil Alpine desert |
|
| Other habitats | Permanent ice and snow |
- a The descriptions of natural habitat types are standard terms in the Habitats Classification Scheme (Version 3.1) developed by the International Union for Conservation of Nature and Natural Resources (IUCN) (IUCN, 2019).
- b The descriptions of land-use/cover types can be found in Liu, et al. (2005a) and W. Song and Deng (2017).
The spatial distribution of natural habitat in China in 2015 is shown in Figure 1. Our study focused on the 31 regions (including provinces, municipalities, and autonomous regions) of mainland China. Forest is mainly distributed in the northeastern and southern areas of China, while savanna and shrubland are mainly distributed in the southwestern areas. Grassland is distributed in the southwestern areas. Meanwhile, wetlands are distributed in the northwestern and eastern areas, and the other types of natural habitat (including rock areas, desert, and other habitats) are intensively distributed in the northwestern areas.
2.1.2 Road datasets
Given the importance of roads with regards to habitat quality (Sharp et al., 2016), four types of road/rail networks were used in this study: national roads, provincial roads, highways, and railways. The vector data of the road datasets came from China's National Basic Geographic Information System (http://nfgis.nsdi.gov.cn).
2.2 Research framework
The research framework consists of two subsequent parts: assessing the impacts of cropland expansion and urban expansion on the quantity (Figure 2) and quality (Figure 3) of natural habitat between 2000 and 2015.
To compare the impacts of cropland expansion and urban expansion on the quantity of natural habitat (Figure 2), first, we calculated spatial distributions of cropland expansion and urban expansion between 2000 and 2015 by adopting map algebra in ArcGIS 10.6 (as described in Section 2.3.1). Then, based on the spatial distributions of cropland expansion and natural habitat, we obtained the changes in spatial distributions of natural habitat due to cropland expansion. The changes in the spatial distribution of natural habitat due to urban expansion were obtained in a similar manner. Finally, with the 'Zonal statistic' tool in ArcGIS 10.6, we calculated the changes in natural habitat due to cropland expansion and urban expansion in each region, respectively (as described in Section 2.3.3).
We assessed the changes in habitat quality due to cropland expansion and urban expansion based on the following steps (Figure 3). First, the land-use map of cropland expansion was generated based on the spatial distribution of cropland expansion between 2000 and 2015 and the land-use/cover map in 2000. The land-use map of urban expansion was also obtained in a similar manner. Then, the habitat quality map in 2000 was calculated, by inputting the land-use/cover map in 2000 and the required parameters (including threat sources, habitat score, and sensitivity) to the InVEST Habitat Quality model. Similarly, the habitat quality maps of 2015, under the impacts of cropland expansion and urban expansion, were respectively obtained and these maps were compared with the habitat quality map in 2000. Thus, the distribution maps of changes in habitat quality due to cropland expansion and urban expansion were obtained. Finally, we classified grades for changes in habitat quality and calculated the area at each grade in each region by the 'Zonal statistic' tool in ArcGIS 10.6 (as described in Section 2.3.3).
2.3 Method
2.3.1 Map algebra
and 
. Subsequently, the 'Raster Calculator' tool was adopted to obtain the transfer map of these types TMt1,t2 accordingly (W. Chen et al., 2020):
(1)Based on the transfer map, TMt1,t2 (with the format of raster), we obtained the transferring matrix that reflects the quantitative transferring relation between different types. Given that the main aim of this study is to explore the impacts of cropland expansion/urban expansion on the natural habitat, we focused on the natural habitat converted to cropland/urban land. Then, the area of natural habitat converted to cropland/urban land in each region was calculated by the 'Zonal statistics' tool, respectively (as described in Section 2.3.3).
2.3.2 InVEST habitat quality model
The InVEST Habitat Quality model has the advantages of the simple requirement of input parameters and spatial visualization of the results. Thus, the model is a widely used assessment model to calculate habitat quality in case-studies around the world (Moreira, Fonseca, Vergílio, Calado, & Gil, 2018; Sharp et al., 2016; Sun, Jiang, Liu, & Zhang, 2019). The spatial distribution of habitat quality can be easily obtained by inputting raster maps (e.g., land-use/land cover map, and habitat distribution map) and required parameters, including threat sources, habitat score, and sensitivity (InVEST User's Guide, 2020; Sharp et al., 2016).
(2)Selecting the threat sources is a key issue when calculating the threat level in the model. Given that China experiences rapid developments in agricultural activities, urbanization, and industrialization, cropland, urban land, and rural construction land were selected as the major threats for habitat. The road/rail networks were also considered as the threats for habitat, including national roads, provincial roads, highways, and railways. The maximum distance and weight of the threats affecting habitat quality were set based on the cognate studies focused on China (Sun et al., 2019; J. Wu, Cao, Shi, Huang, & Lu, 2015) (Table 2).
| Threat factors | Maximum distance of influence/km | Weight | Type of decay over space | References |
|---|---|---|---|---|
| Cropland | 7 | 0.8 | Exponential | Sun et al. (2019) |
| Urban land | 10 | 0.95 | Exponential | Sun et al. (2019) |
| Rural construction land | 9 | 0.9 | Exponential | Sun et al. (2019) |
| National road | 10 | 0.6 | Linear | J. Wu et al. (2015) |
| Provincial road | 5 | 0.4 | Linear | J. Wu et al. (2015) |
| Highways | 8 | 0.5 | Linear | J. Wu et al. (2015) |
| Railways | 7 | 0.5 | Linear | J. Wu et al. (2015) |
The parameters of habitat suitability in Table 3 were based on the studies of Sun et al. (2019) and Xie, Zhen, Lu, Yu, and Cao (2008). The sensitivity of habitat types to each threat was based on the published literature focused on China (Y. Deng et al., 2018; Z. Liu, Tang, et al., 2017b; C. Liu & Wang, 2018; Sun et al., 2019; J. Wu et al., 2015; Zhong & Wang, 2017) (Table 3).
| Natural habitat types | Habitat score | Threat factors | References | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Cropland | Urban land | Rural construction land | National roads | Provincial roads | Highways | Railways | |||
| Forest | 0.95 | 0.6 | 0.9 | 0.85 | 0.9 | 0.7 | 0.7 | 0.8 | Sun et al. (2019); J. Wu et al. (2015); Zhong & Wang (2017) |
| Savanna | 0.5 | 0.2 | 0.6 | 0.55 | 0.8 | 0.8 | 0.8 | 0.7 | Sun et al. (2019); C. Liu & Wang (2018) |
| Shrubland | 0.6 | 0.6 | 0.65 | 0.6 | 0.7 | 0.5 | 0.5 | 0.6 | Sun et al. (2019); C. Liu& Wang (2018); Zhong & Wang ( 2017) |
| Grassland | 0.6 | 0.4 | 0.47 | 0.8 | 0.9 | 0.7 | 0.7 | 0.8 | Sun et al. (2019); C. Liu & Wang(2018); Z. Liu, Tang, et al. (2017b); Y. Deng, Jiang, Wang, Lu, and Chen (2018) |
| Wetlands | 1 | 0.7 | 0.7 | 0.65 | 0.6 | 0.4 | 0.5 | 0.5 | Sun et al. (2019); J. Wu et al. (2015); Zhong & Wang (2017) |
| Rocky areas | 0.1 | 0.1 | 0.3 | 0.3 | 0.3 | 0.3 | 0.3 | 0.3 | Zhong & Wang (2017) |
| Desert | 0.1 | 0.1 | 0.3 | 0.3 | 0.3 | 0.3 | 0.3 | 0.3 | Zhong & Wang (2017) |
| Other habitats | 0.1 | 0.1 | 0.3 | 0.3 | 0.3 | 0.3 | 0.3 | 0.3 | Zhong & Wang (2017) |
Habitat degradation refers to the set of processes by which habitat quality is reduced (Terrado, Sabater, & Acuña, 2016). Since habitat quality ranges from 0 to 1, habitat degradation correspondingly varies from 0 to 1 (Sharp et al., 2016). Meanwhile, habitat degradation is heavy if its degradation is over 0.5, indicating the decline in habitat quality is at least 0.5 (Bai, Xiu, Feng, & Liu, 2019; J. He, Huang, & Li, 2017b). Thus, we were concerned more about the natural habitat that suffered heavy degradation in this study, and habitat degradation was further classified into five grades: grade 1 (0.9–1), grade 2 (0.8–0.9), grade 3 (0.7–0.8), grade 4 (0.6–0.7), and grade 5 (0.5–0.6). The area of habitat degradation in different grades in each region as a result of cropland expansion and urban expansion was calculated by the 'Zonal statistics' tool, respectively (as described in Section 2.3.3).
2.3.3 'Zonal statistics'
The 'Zonal statistics' tool in the ArcGIS software platform is a widely used tool to conduct spatial analysis (Dong, Sadeghinaeenifard, Xia, & Tan, 2019; Goodman, BenYishay, Lv, & Runfola, 2019). With the 'Zonal statistics' tool, a statistic is calculated for each zone defined by a zone dataset, based on the values from another dataset (a value raster) (ESRI, 2020). Namely, zonal statistics refer to identifying measurement data (i.e., pixels from raster data) relevant to a given boundary feature and aggregating measurement values using a specified aggregation method (e.g., sum, mean, min, and max) (Goodman et al., 2019). In this study, the 'Zonal statistics' tool was applied to calculate the quantity of habitat loss and the area of habitat degradation in each region, as a result of cropland expansion and urban expansion.
To calculate the quantity of habitat loss in each region as a result of cropland expansion and urban expansion, firstly, the Chinese mainland administration zones map was set as the boundary feature, while the transfer map of natural habitat to cropland/urban land was set as the input raster layer (containing the values on which to calculate a statistic). Then, the numbers of raster in which natural habitat converted to cropland/urban land in each region were calculated by the 'Zonal statistics' tool. Finally, the converted area of each habitat type in each region was obtained since one raster cell is 1 km2.
To assess the area of habitat degradation in each region, firstly, the raster maps of habitat degradation in different grades were obtained (as described in Sections 2.2 and 2.3.2), respectively. Then, the raster map of habitat degradation in a specific grade (including grade 1–5) was set as the input raster layer, while the Chinese mainland administration zones map was set as the boundary feature. Finally, the numbers of the raster of habitat degradation at different grades in each region were obtained by the 'Zonal statistics' tool, and their corresponding area was also calculated.
3 RESULTS
3.1 Loss in quantity of natural habitat due to cropland expansion and urban expansion
3.1.1 Natural habitat loss due to cropland expansion
The total amount of natural habitat lost in China between 2000 and 2015 as a result of cropland expansion was 35,811 km2. Grassland lost the largest area among all types of natural habitat with 18,175 km2, which accounted for 51% of the total natural habitat lost, followed by wetlands (Table 4). Comparatively, the loss of wetlands was about half of the loss of grassland, while the loss of desert was one-quarter of that loss. The losses of forest and shrubland were almost equal. There was no loss of other habitats.
| Natural habitat types | Impacts of cropland expansion | Impacts of urban expansion | Ratiob | ||
|---|---|---|---|---|---|
| Natural habitat loss (km2) | Percentagea | Natural habitat loss (km2) | Percentagea | ||
| Forest | 2,158 | 6% | 828 | 28% | 3 |
| Savanna | 921 | 3% | 326 | 11% | 3 |
| Shrubland | 1,817 | 5% | 120 | 4% | 15 |
| Grassland | 18,175 | 51% | 871 | 29% | 21 |
| Wetlands | 7,860 | 22% | 685 | 23% | 11 |
| Rocky areas | 60 | 0% | 5 | 0% | 12 |
| Desert | 4,820 | 13% | 147 | 5% | 33 |
| Other habitats | 0 | 0% | 0 | 0% | — |
| Total | 35,811 | 100% | 2,982 | 100% | 12 |
- a The numeric percentage indicates the percentage of loss in each type of natural habitat.
- b Ratio = Natural habitat loss due to cropland expansion/Natural habitat loss due to urban expansion.
There were apparent spatial differences in both the losses of each type of natural habitat (Figure 4a–g) and the total loss of all types of natural habitat (Figure 4h) caused by cropland expansion. Figure 4a shows that the large loss of forest was mainly distributed orthern China. The largest loss of forest was 777 km2 in Heilongjiang, accounting for 36% of the total loss of forest in the whole of China. Meanwhile, the losses of savanna, shrubland, grassland, and wetlands declined from the northwest to the southeast (Figure 4b–e). The largest losses of these habitat types (except for shrubland) all occurred in Xinjiang with 324 km2 (accounting for 35% of the total loss of savanna), 11,399 km2 (63% of total loss of grassland), and 2,681 km2 (34% of total loss of wetlands), respectively. Comparatively, the spatial distributions of the losses of rocky areas (Figure 4f) and desert (Figure 4g) were similar: the losses of these habitat types mostly occurred in the northern regions, while there was no loss in the remaining regions. It is worth noting that the loss of each type of natural habitat from cropland expansion was mostly distributed in northern China. Furthermore, the largest loss of almost all types of natural habitat (except for forest, and shrubland), caused by cropland expansion, was in Xinjiang.
The total loss of all types of natural habitat due to cropland expansion showed a declining trend from the north to south in China (Figure 4h). Specifically, the amount of natural habitat lost in Xinjiang was 19,086 km2 (53%), followed by Heilongjiang, Inner Mongolia, Gansu, and Jilin. In contrast, the loss of natural habitat due to cropland expansion was much lower in the southeastern regions. The loss of natural habitat in Shanghai, Fujian, and Guangdong all was less than 50 km2.
The conversion of each natural habitat to cropland showed spatial heterogeneity (Figure 4i). Forest dominated in the southeastern regions, while wetlands dominated in the northeastern regions. Comparatively, grassland dominated in both the northwestern and southwestern regions.
3.1.2 Natural habitat loss due to urban expansion
The total amount of natural habitat loss was 2,982 km2 in China between 2000 and 2015 as a result of urban expansion. Grassland lost most among all types of natural habitat with 871 km2, accounting for 29% of the total loss, followed by forest (28%) and wetlands (23%) (Table 4). Comparatively, savanna loss was one-third of grassland loss, while shrubland and desert lost much less. There was no loss of other habitats.
The spatial distributions of losses of forest, savanna, and shrubland as a result of urban expansion were similar, and these losses generally increased from the northwest to the southeast (Figure 5a–c). Meanwhile, the largest losses of these habitat types all occurred in Guangdong, with 362 km2 (accounting for 44% of the total loss of forest), 66 km2 (20% of total loss of savanna), and 21 km2 (18% of total loss of shrubland), respectively. In contrast, the loss of grassland declined from the northwest to the southeast (Figure 5d). The largest losses of grassland were in Inner Mongolia and Xinjiang, accounting for 54% of total loss of grassland. Comparatively, the loss of wetlands was mainly in eastern China, while the loss in each of the rest of the regions was less than 10 km2 (Figure 5e). The loss of rocky areas only occurred in Xinjiang and Jilin (Figure 5f). Comparatively, the loss of desert mainly occurred in northern China, while there was almost no loss in southern China (Figure 5g). Noticeably, the spatial distributions of the losses of each type of natural habitat due to urban expansion were very different; the losses of forest, savanna, shrubland, and wetlands mostly occurred in southeastern China, while the losses of the remaining types of habitat mostly occurred in northern China.
The total loss of natural habitat due to urban expansion generally decreased from the southeast to the northwest (Figure 5h). The largest loss of natural habitat due to urban expansion was in Guangdong with 759 km2, accounting for 25% of the total natural habitat loss caused by urban expansion in China during this period.
Overall the types of natural habitat lost due to urban expansion varied among the different regions (Figure 5i). The main losses were forest and wetlands in southeastern and northeastern China, while the primary loss was grassland both in northwestern and southwestern China.
3.1.3 A comparison of the impacts of cropland expansion and urban expansion on natural habitat loss
Both the total amount of natural habitat loss and the amount of each type of natural habitat loss due to cropland expansion far exceeded the losses due to urban expansion (Table 4). It is worth noting that the total loss of natural habitat due to cropland expansion was twelve-times as much as that from urban expansion. In addition, cropland expansion had the largest impact on grasslands and wetlands while urban expansion resulted in the largest impacts on grassland and forest.
The ratio of the total amount of natural habitat lost due to cropland expansion to that lost due to urban expansion was calculated (Figure 6). The ratio above 1 represents the total amount of natural habitat lost due to cropland expansion exceeds that due to urban expansion. In contrast, the ratio below 1 represents the total amount of natural habitat lost due to urban expansion exceeds that due to cropland expansion. The ratio equal to 1 represents the total amount of natural habitat lost due to cropland expansion equal to that due to urban expansion. Generally, the total amount of natural habitat lost due to cropland expansion exceeded that of urban expansion in most regions of China, especially in the northwestern and northeastern parts. The highest ratio was 177 in Heilongjiang, followed by Xinjiang (54), and Jilin (28). However, it was the opposite of some developed regions of China. The impacts of cropland expansion on natural habitat loss were lower than that of urban expansion in Beijing (0.31), Fujian (0.25), and Guangdong (0.06). Comparatively, the impacts of cropland expansion and urban expansion on natural habitat loss were equal in Sichuan.
3.2 Habitat degradation due to cropland expansion and urban expansion
3.2.1 Habitat degradation due to cropland expansion
The total area of habitat degradation due to cropland expansion was 29,829 km2 (Table 5). The area of habitat degradation was greatest in grade 5, accounting for 66% of the total area, followed by grade 1 (32%). The area of habitat degradation in grade 2 and 4 was almost equal, taking up 1% of the total area, respectively.
| Habitat degradation grades | Impacts of cropland expansion | Impacts of urban expansion | Ratiob | ||
|---|---|---|---|---|---|
| Area of habitat degradation (km2) | Percentagea | Area of habitat degradation (km2) | Percentagea | ||
| Grade 1 (0.9–1) | 9,530 | 32% | 1,169 | 47% | 8 |
| Grade 2 (0.8–0.9) | 215 | 1% | 175 | 7% | 1 |
| Grade 3 (0.7–0.8) | 103 | 0% | 71 | 3% | 1 |
| Grade 4 (0.6–0.7) | 200 | 1% | 50 | 2% | 4 |
| Grade 5 (0.5–0.6) | 19,781 | 66% | 1,000 | 41% | 20 |
| Total | 29,829 | 100% | 2,465 | 100% | 12 |
- a The numeric percentage indicates the percentage of the area of habitat degradation at each grade.
- b Ratio = Area of habitat degradation due to cropland expansion/Area of habitat degradation due to urban expansion.
The habitat degradation in grade 1 was mostly distributed in northeastern and northwestern China (Figure 7a). The area of habitat degradation in grade 1 was largest in Heilongjiang (2,947 km2) and Xinjiang (2,897 km2), accounting for 61% of the total degraded area at this grade. Spatial distributions of the habitat degradation in grade 2 (Figure 7b), grade 3 (Figure 7c), and grade 4 (Figure 7d) were similar: habitat degradation was mostly distributed in northeastern China, while there was almost no degradation in the remaining regions. Comparatively, the natural habitat of all regions except for Beijing suffered degradation in grade 5 (Figure 7e). The total area of habitat degradation decreased from the northwest to the southeast (Figure 7f). The total area of habitat degradation was the largest in Xinjiang with 15,029 km2 (50% of the degraded areas in the whole of China).
3.2.2 Habitat degradation due to urban expansion
The total area of habitat degradation due to urban expansion was 2,465 km2 (Table 5). Most natural habitat suffered degradation in grade 1 (47% of the total area) and in grade 5 (41% of the total area). Comparatively, the area of habitat degradation at the remaining grades was much less. The area of habitat degradation in grade 2 took up 7% of the total area, while that in grade 3 and 4 accounted for 3% and 2%, respectively.
The habitat degradation in grade 1 was mostly distributed in southeastern China (Figure 8a). The largest area of habitat degradation in grade 1 was in Guangdong with 421 km2, accounting for 36% of the total degraded area at this grade. Comparatively, the habitat degradation in grade 2 (Figure 8b), grade 3 (Figure 8c), and grade 4 (Figure 8d) were mostly distributed in the southeastern regions. Meanwhile, the natural habitat of all regions except for Shanghai suffered degradation in grade 5 (Figure 8e). The total area of habitat degradation decreased from the southeast to the northwest (Figure 8f).
3.2.3 A comparison of the area of habitat degradation due to cropland expansion with urban expansion
The total area of habitat degradation due to cropland expansion (29,829 km2) far exceeded that due to urban expansion (2,465 km2), about twelve-times (Table 5). Furthermore, at each grade of habitat degradation, the area of habitat degradation due to cropland expansion also exceeded that due to urban expansion. Noticeably, the area of the heaviest habitat degradation (namely, in grade 1) due to cropland expansion was 9,530 km2, which was eight times as much as that area due to urban expansion (1,169 km2). Meanwhile, cropland expansion had more severe impacts on habitat degradation in grade 5 while urban expansion comparatively had more severe impacts on habitat degradation in grade 1.
The ratio of the total area of habitat degradation due to cropland expansion to that total area due to urban expansion was calculated (Figure 9). The ratio above 1 represents the total area of habitat degradation due to cropland expansion exceeds that due to urban expansion. In contrast, the ratio below 1 represents the total area of habitat degradation due to urban expansion exceeds that due to cropland expansion. Generally, the area of habitat degradation due to cropland expansion exceeded that of urban expansion in most regions of China. The ratio showed an increasing tendency from southeast to northwest. The highest ratio was 166 in Heilongjiang, followed by Xinjiang (63). In contrast, in the southeastern regions, the impacts of cropland expansion on the habitat degradation were lower than that of urban expansion, such as Fujian (0.28) and Guangdong (0.06).
In summary, the impacts of cropland expansion on both the natural habitat loss and degradation far exceeded the impacts of urban expansion. Meanwhile, the spatial distributions of such impacts of cropland expansion and urban expansion were completely different. The natural habitat loss and degradation as a result of cropland expansion was decreased from the northwest to the southeast. It is noteworthy that the largest natural habitat loss and area of degradation as a result of cropland expansion was in Xinjiang and Heilongjiang. However, the natural habitat loss and degradation as a result of urban expansion showed the opposite pattern: decreasing from the southeast to the northwest. The largest natural habitat loss and area of degradation as a result of urban expansion was in southeastern China, especially in Guangdong.
4 DISCUSSION
4.1 Considerable impacts of cropland expansion on both the natural habitat loss and degradation
Existing studies have explored the substantial impacts of cropland expansion on natural habitat loss and degradation (Molotoks et al., 2018; Sun et al., 2019; van Vliet, 2019), while such impacts also were assessed in this study. Furthermore, we assessed how serious the impacts of cropland expansion are on both the natural habitat loss and degradation at the national scale, by comparing the impacts from cropland expansion and urban expansion. The results demonstrated that both the natural habitat loss and degradation, caused by cropland expansion, were more serious than urban expansion. Consequently, a considerable natural habitat loss and degradation, caused by cropland expansion, need to be more concerned.
In addition, combing the spatial distributions of habitat loss and degradation, we found that the greatest impacts of cropland expansion on both habitat loss and degradation mostly occurred in areas where the ecological environment is already vulnerable (namely, ecologically vulnerable regions) (J. Liu et al., 2015; Zhao, Ji, Tian, Chen, & Wang, 2018), whereas the impacts of urban expansion were much less in these areas. In detail, ecologically vulnerable regions refer to the regions where the resistance and resilience of ecosystems in response to external interference are weak (Adger, 2006; Beroya-Eitner, 2016), and are prone to ecological degradation (J. Liu et al., 2015; Zou & Yoshino, 2017). J. Liu et al. (2015) identified 18 ecologically vulnerable regions mostly distributed in the northwestern, northeastern, and southwestern regions of China. Meanwhile, Zhao et al. (2018) suggested that the vulnerability degree in western China is significantly higher than in eastern China. Furthermore, western China presents a large contiguous area of a higher degree of environmental vulnerability, such as Xinjiang, Tibet, and Qinghai (Zhao et al., 2018). Thus, it is critical to conserve the natural habitat in these ecologically vulnerable regions to avoid further destruction.
As demonstrated in this study, cropland expansion mostly occurred in the ecologically vulnerable regions (Liu et al., 2015; Zhao et al., 2018). This is related to that most ecologically vulnerable regions are poverty-stricken regions in China, and residents in such regions highly depend on agriculture. In comparison, given that the overall ecological environment in eastern China is superior, this supports the conditions for rapid economic development, plus it has the advantage of having a favorable location conditions (X. Deng, Huang, Rozelle, Zhang, & Li, 2015). In this case, the demand for urban land rapidly increases to meet the economic development (C. He et al., 2014).
This study highlights the considerable loss in both the natural habitat loss and degradation caused by cropland expansion. The result can help enhance our understanding of the serious impacts of cropland expansion on natural habitat. It can also draw our attention to the natural habitat in areas undergoing extensive cropland expansion and trigger us to take action for protecting natural habitat in these areas.
4.2 Explanations for differences in impacts of cropland expansion and urban expansion on natural habitat
The spatial differences in natural habitat loss due to cropland and urban expansion can be explained by land-use displacement (the geographic displacement of land-use activities) (van Vliet, 2019). Urban expansion has taken large amounts of cropland, while cropland reclamation resulted in the loss of natural habitat to compensate for the loss of cropland within or across regions (Tang, Ke, Zhou, Zheng, et al., 2020b; van Vliet, 2019). Therefore, cropland displacement occurs. This is true both in worldwide (van Vliet, 2019) and China (Zuo et al., 2018). In China, most cropland is replaced by urban land in southeastern regions, while cropland reclamation is mostly distributed in northwestern and northeastern regions (Zuo et al., 2018).
The differences among the natural habitat types converted to cropland or urban land in different regions are mainly related to the prevailing types of natural habitat in different regions (van Vliet, 2019). Comparatively, the reasons for the differences in the natural habitat loss and degradation due to cropland expansion among the different regions include the natural and socioeconomic factors. The natural factors mainly refer to the amount of natural habitat and climate change (Iizumi & Ramankutty, 2015), while the socioeconomic factors mainly refer to the status of agriculture (Hertel, Ramankutty, & Baldos, 2014), cropland protection policies (Popp et al., 2017; W. Song & Liu, 2017; X. Song, Ouyang, Li, & Li, 2012), and environmental protection policies (Bai et al., 2019). In contrast, the main reasons for the differences in the natural habitat loss and degradation due to urban expansion among the different regions are the socioeconomic factors, such as the level of urbanization and economic development (Wang, Li, & Fang, 2018), and urbanization policies (Kuang et al., 2016; Y. Yang, Liu, & Zhang, 2017). For example, the southern regions of China have rapid urbanization and economic development, and, meanwhile, natural habitat is replaced by urban land in these regions (C. He et al., 2014).
4.3 Potential impacts of natural habitat loss and degradation on ecosystem services
Considering the vulnerability and irreplaceability of natural habitat, the impacts of natural habitat loss and degradation on ecosystem services have been explored, especially biodiversity (Di Marco et al., 2019; Molotoks et al., 2018; Powers & Jetz, 2019). Based on the synthesis of fragmentation experiments spanning multiple biomes and scales, five continents, and 35 years (approximately started from 1980), Haddad et al. (2015) found that habitat fragmentation had reduced biodiversity by 13 to 75%. 1,700 species of amphibians, birds, and mammals are expected to become endangered due to natural habitat transformations, based on the global decadal land-use projections to the year 2070 (Powers & Jetz, 2019). The number of plant species committed to extinction over the long term increased by 60% globally between 1900 and 2015 (from 10,000 to 16,000), and this number is projected to substantially increase (to 18,000) by 2050 (Di Marco et al., 2019). Furthermore, natural habitat loss and degradation affect species differently, depending on how sensitive they are to the change of natural habitat (Keinath et al., 2017; Todd, Rose, Price, & Dorcas, 2016). More sensitive species are often at greater risk of decline from natural habitat loss and degradation than less-sensitive species (Todd et al., 2016). Quesnelle, Lindsay, and Fahrig (2014) found that species with lower mobility and/or reproductive rates were more threatened since their sensitivity to habitat loss is higher than species with greater mobility and/or higher reproductive rates. Meanwhile, Keinath et al. (2017) indicated that amphibians showed lower sensitivity than birds, mammals, and reptiles.
In China, the threatened mammals are mostly distributed in southwest China, where there are 10–40 types threatened per 100 km2; the threatened birds are mostly distributed in northeast, northwest, and southeast China, where there are over 10 types threatened per 100 km2; and the threatened amphibians are mostly distributed in Hainan and in some southwestern regions, where there are over six types threatened per 100 km2 (Mapping the World's Biodiversity, 2019; Zhang et al., 2016). Given that habitat loss and degradation directly affect the environmental conditions for species (Malavasi, Bartak, Carranza, Simova, & Acosta, 2018), based on the spatial distributions of habitat loss and degradation caused by cropland expansion and urban expansion in this study, the threatened mammals and amphibians might be possibly affected more by cropland expansion, while threatened birds might be likely affected by both cropland expansion and urban expansion.
Natural habitat loss also has posed a great and immediate threat to other ecosystem services, such as carbon storage, soil retention, and sandstorm prevention (Mao et al., 2019; Molotoks et al., 2018; Tang, Ke, Zhou, Zheng, et al., 2020b). Cropland expansion was expected to lead to biomass carbon storage and soil carbon storage lost with 13.7 and 4.6% between 2010 and 2050, respectively (Molotoks et al., 2018). Meanwhile, global urban expansion was projected to result in a carbon storage loss of 1.38 Pg C between 2000 and 2030 (Seto et al., 2012). In addition, Mao et al. (2019) indicated that cropland expansion was the dominant factor driving the losses of forest, grassland, and wetlands in northeastern China between 2000 and 2015, which caused soil retention and carbon storage to decrease.
4.4 Main drivers of cropland expansion
The drivers of cropland expansion have been explored both at the global (Molotoks et al., 2018; van Vliet, 2019) and regional scales (C. Chen, Zhang, & Lu, 2016; Y. Zhang, Li, Wang, Cai, & Bao, 2015), such as socioeconomic, natural, political, technological, and cultural drivers (Newman, McLaren, & Wilson, 2014). The world's growing food demands (Foley et al., 2011) are regarded as a key socioeconomic driver (Molotoks et al., 2018). Global crop demand was predicted to increase by 100–110% from 2005 to 2050 with an increasing global population (Tilman, Balzer, Hill, & Befort, 2011). While increased production can partially be met by improving cropland intensification, it is limited (Popp et al., 2017). In this case, global cropland is expected to expand to meet the demand for global food in the future. Another crucial driver of cropland expansion is land-use policies (Debonne, van Vliet, & Verburg, 2019; Shen et al., 2017; W. Song & Pijanowski, 2014). Interventionist, preventive, and noninterventionist policies when conduct large-scale land acquisitions will lead to different cropland expansion patterns in Cambodia (Debonne et al., 2019). Cropland use intensity (Byerlee, Stevenson, & Villoria, 2014; W. Wu et al., 2018), agricultural markets (Hertel et al., 2014), and population and foreign labor migration (Paudel et al., 2016) also trigger cropland expands.
Cropland expansion in China is affected by natural drivers, especially climate change (Shi et al., 2014). Climate warming has aided the cropland expansion in northern China (Shi et al., 2014). Moreover, cropland expansion also has a complex linkage to socioeconomic drivers, such as government policies (Shi et al., 2014) and the improvement of agricultural technology (Y. Zhang et al., 2015). With cropland use policies, Chinese local jurisdictions continued expropriating cropland on a large-scale for development (Shen et al., 2017). Y. Zhang et al. (2015) indicated that the main drivers of cropland expansion in Xinjiang, China, include improvements in agricultural infrastructure and production technology. Cropland expansion also depends on the decision of farmers (X. Liu, Wang, Liu, & Meng, 2005b). For example, if farmers want to increase income it is more likely to expand cropland (X. Liu, Wang, et al., 2005b).
4.5 Implications for Chinese cropland protection policies and initiatives
Cropland expansion leads to a large loss of natural habitat, which indicates that we should rethink the Chinese cropland protection policies. China's Requisition–Compensation Balance of Cropland Policy clarifies the responsibilities and obligations of cropland requisition–compensation (Y. Liu, Fang, & Li, 2014; Shen et al., 2017). This policy requires new cropland to compensate for the lost cropland due to urban expansion with the land of the same quantity (the stakeholders occupying cropland are responsible for the compensation equivalent to the requisitioned cropland) (“Decision of the State Council on Deepening Reform and Strict Land Management”, 2004; Y. Chen et al., 2018; Yu et al., 2018). Another policy closely related to the cropland expansion is the General Dynamic Balance of Cropland (X. Liu, Zhao, & Song, 2017a; Y. Wu, Shan, Guo, & Peng, 2017). This policy was launched by the Chinese Central Government, which aims to maintain the quantity and quality of cropland and reverse the trend of a reduction in cropland (X. Liu, Zhao, et al., 2017a). Cropland expansion with these protect policies has effectively maintained cropland resources that are the basis of achieving national food demand and helped to increase crop production by 66% from 1987 to 2010 (Zuo et al., 2018). Although cropland expansion has positive impacts on crop production, it inevitably leads to considerable natural habitat loss.
In contrast, some of the cropland protection initiatives in China can mitigate the natural habitat loss caused by cropland expansion, since these initiatives aim to enhance the quality of cropland to guarantee food security rather than only maintain the quantity of cropland. The Chinese Government emphasizes that efforts should be made to strengthen the comprehensive protection of the quantity, quality, and ecological functions of cropland (Liu et al., 2017a). This comprehensive protection aims to comprehensively maintain and improve the production functions, ecological functions, and living functions of cropland ecosystems, and to guarantee regional and national food security, ecological security, and social stability. Furthermore, the Chinese Government protects the basic cropland of high quality (basic cropland is the cropland that cannot be occupied according to the “Agricultural Law of the People's Republic of China” and “Land Administration Law of the People's Republic of China”). Sustainable land remediation is also an effective way to conserve the quality and ecological functions of cropland since it can bring about beneficial reuse capabilities for the land and create a net environmental benefit (Hou et al., 2018).
Therefore, it is necessary to consider the loss of natural habitat due to cropland expansion when implementing cropland protection policies. It is also significant to improve the cropland protection policies and initiatives to guarantee food security with little or no harm to the natural habitat (Cazalis, Loreau, & Henderson, 2018).
4.6 Limitations
The limitation of our study is that only the observed land-use data for China from specified time points (2000 and 2015) were adopted. Time series of land-use data should be used to explore the dynamic process of natural habitat loss due to cropland expansion. In addition, the threats on the landscape are additive in the InVEST Habitat Quality model, whereas the collective impact of multiple threats is much greater than the sum of individual threat levels in some cases (Sharp et al., 2016). Meanwhile, the parameters used in the InVEST Habitat Quality model to assess habitat quality in this study were from the published literature instead of the sampling method or a biogeochemical method. There are many biophysical or anthropogenic factors that act as degradation sources or as threats to habitat, such as alien species invasion, while we only considered factors from the perspective of land-use change, including cropland expansion and urban expansion. It should be acknowledged that regional natural habitat is also impacted by threats from the wider landscape outside country boundaries, so it may be more comprehensive to incorporate such impact factors. Moreover, the InVEST Habitat Quality model is a type of assessment model rather than a simulation model or interpolation model. So, there is no validating process in the InVEST model.
5 CONCLUSIONS
This study explored the magnitude of the impacts of cropland expansion on both the natural habitat loss and degradation in China between 2000 and 2015 by comparing them with the impacts of urban expansion. The impacts of cropland expansion on both the natural habitat loss and degradation were higher than that of urban expansion. Importantly, the impacts of cropland expansion on both the natural habitat loss and degradation mostly occurred in areas where the ecological environment is already vulnerable, whereas the impacts of urban expansion were much less in these areas. Therefore, our results suggest that it is significant to concern the impacts of cropland expansion on natural habitat and to take these impacts into consideration when making and implementing the cropland protection policies related to cropland expansion.
ACKNOWLEDGMENTS
This work was financially supported by the National Natural Science Foundation of China (Grant Nos. 41971240, and 41371113), Post-finance Project for National Social Sciences (Grant No. 19FGLB071), Fundamental Research Funds for the Central Universities (Grant No. 2662017PY063), and Post-finance Project for Philosophy and Social Sciences of the Ministry of Education (Grant No. 18JHQ081). The authors thank the Proof-Reading-Service Web Shop (https://www.proof-reading-service.com) for its linguistic assistance for this manuscript. The authors appreciate that editors and anonymous reviewers proposed insightful comments to help us improve this paper.
Open Research
DATA AVAILABILITY STATEMENT
The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.
