Spatio-temporal variation and future multi-scenario simulation of carbon storage in Bailong River Basin using GeoSOS-FLUS and InVEST models

doi:10.21203/rs.3.rs-3138310/v1

Download PDF

Research Article

Spatio-temporal variation and future multi-scenario simulation of carbon storage in Bailong River Basin using GeoSOS-FLUS and InVEST models

https://doi.org/10.21203/rs.3.rs-3138310/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

By providing a scientific foundation for managing regional ecosystem carbon (C) pools, research on the spatial distribution characteristics of regional C stocks can assist in the development of policies on C emissions reduction and sequestration enhancement. Using the GeoSOS-FLUS and InVEST models and explorations of the Bailong River Basin in the past 20 years, the influence of three future scenarios of land use change—natural development (ND), ecological protection (EP) and arable land protection (ALP)—on C storage was modelled. Between 2000 and 2020, there was a gradual increase in C storage in the BRB with a total increase of 5.58 Tg (3.19%), showing notable spatial heterogeneity. The increase in C storage was attributed to land use conversion among woodland, arable land and grassland, with the conversion between woodland and arable land being the primary factor contributing to the increase in C storage. By 2050, C storage under the EP, ALP and NP scenarios was 183.915, 183.108 and 183.228 Tg, respectively. In 2050, C storage under the EP scenario increased by 0.37% compared with that in 2020, and decreased by 0.07% and 0.005% under the ALP and NP scenarios, respectively. In contrast to the other scenarios, the EP scenario prioritised the protection of the woodland and grassland C sinks, which has significant implications for future planning.

Carbon storage

InVEST model

GeoSOS-FLUS model

Land use

Bailong River Basin

Terrestrial ecosystem sequestration of carbon (C) is now considered to be one of the most economical, practical and sustainable approaches to reducing atmospheric carbon dioxide (CO₂) concentrations, and is generally accepted internationally (Fang et al. 2015). Studies have shown that there is a close correlation between C storage and land use/cover change (LUCC) (W. Li et al. 2022). LUCC significantly influences C stocks on a regional scale by changing the amount of C stored in vegetation and soil within the ecosystem. This change often results in significant C exchange between the ecosystem and the atmosphere (Vizcaíno-Bravo et al. 2020; Yang et al. 2021). According to the IPCC report, LUCC leads to the release of C into the atmosphere at a rate of 1.5 Pg⋅a^–1, reflecting the significant impact of LUCC on the global C cycle of terrestrial ecosystems (He et al. 2018). Thus, evaluating the influence of LUCC on ecosystem C stocks and forecasting future C storage is of great significance for the management of C pools and policy formulation.

Carbon in land ecosystems is primarily composed of above- and below-ground material, including soil and dead organic matter. Thus, the vegetation distribution and soil environment will directly affect C storage in land ecosystems(Yin and Porporato 2023). LUCC can also alter the process of C cycling by changing the structure and function of ecosystems (P. Wei et al. 2021). Numerous studies have been conducted on C stock changes in single terrestrial ecosystems (Zhang et al. 2021), such as woodland, farmland, grassland, etc. There has been a focus on C stocks within forests (Mumcu Kucuker 2020; Tian et al. 2022), at forest edges and in soil (Meeussen et al. 2021; Remy et al. 2016) and within single ecosystems (Castellano et al. 2022; L. Deng et al. 2016). There has been less focus on total C stocks in terrestrial ecosystems. In recent decades, the escalation of human activity and climate change have affected plant yields and soil respiration, resulting in significant alterations to ecosystem structure and cycle mechanisms (Yan et al. 2022). Therefore, the trend of ecosystem C change is uncertain. Research ranging from the analysis of single and small-scale ecosystems to large-scale terrestrial ecosystems is essential for effective ecosystem conservation and use.

Field investigations and empirical statistics can provide accurate data on the C density of terrestrial ecosystems (Mattsson et al. 2016; Sahu et al. 2021) and models can be constructed to quantitatively distinguish different factors (Gomes et al. 2019). However, these methods cannot accurately reflect the response of long time series and large-scale C storage to natural and anthropogenic changes (Y. Liu et al. 2021). The integrated valuation of ecosystem services and tradeoffs (InVEST) model has the advantages of spatial representation, dynamic evaluation and quantitative assessment, and can be applied to the assessment of the effect of LUCC on ecosystem C stocks (S. Y. Liu et al. 2018). Many studies have utilised the InVEST model to evaluate the status of C reserves in regional ecosystems. For example, Wei et al. (2021) performed a systematic study on C stocks in the northeastern Qinghai-Tibet Plateau (QTP) (P. Wei et al. 2021). However, this only reflected the changes in regional ecological C stocks and it was lack of basis for predicting future trends in C stocks or the future management of C pools and policy formulation. Therefore, utilising the prediction model to forecast future LUCC and to dynamically monitor and protect the long-term C storage of the ecosystem would be of great value.For land use simulation, the CA-Markov (Hamad et al. 2018; Jana et al. 2022), CLUE-S (Islam et al. 2021; Mamanis et al. 2021), FLUS (Z. Chen et al. 2021) and other models have been extensively utilised. However, these cannot reflect large-scale changes or simulate different scenarios with different influencing factors. However, the GeoSOS-FLUS model imports an adaptive inertia competition mechanism of roulette selection (Hou et al. 2022), which can lead to a more accurate and rational distribution of future land types. Although it has been extensively used for land use change predictions, there have been limited uses of this model in conjunction with the InVEST model to explore ecosystem C stocks.

The Bailong River Basin (BRB) is situated in the eastern QTP from the Loess Plateau to the Qinba Mountains. Containing a typical ecologically fragile belt (Zhang 2016) including high-altitude mountains and a typical ecological ecotone, it is a crucial water source and conservation barrier (Gong et al. 2021). Due to the specific geographical and ecological circumstances of the BRB, vegetation species are abundant and diversity is high. The soil in the region has accumulated rich soil organic material and vegetation has rich C storage, which is a neglected C pool. Additionally, influenced by the climate of the plateau, there is a very significant vertical zonation, forming complex and diverse climate types (Xie 2015). With ongoing global climate changes and population growth, land development is increasing and with powerful tectonic movements in the region, ecosystems are facing enormous challenges, especially in terms of C cycling and C balance. Accordingly, GeoSOS-FLUS and InVEST models were constructed to evaluate the influence of LUCC on C stocks in the BRB, as well as ecosystem C stocks under disparate scenarios. The goal was to provide a scientific understanding of the C stocks in the region for the future management of C pools, policy formulation and sustainable development.

Overview of the study

The BRB is situated in the most vulnerable area of the eastern QTP in south-eastern Gansu Province. It is also a crucial ecological zone for water and soil conservation (Zhang 2016). The BRB mainly extends through Dangchang County, Wudu County, Zhouqu County and the Wenxian region. Its terrain is high in the northwest and low in the southeast and undulating, with a narrow north-west-south-east shuttle-shaped alignment. The valley area is dry and hot with low rainfall and the alpine area is cold and rainy, with a significant distinction between dry and wet seasons. For the mountains in this region, the temperature decreases as the altitude increases (Xie 2015). There are high slopes and cliffs along the river basin, resulting in mountain microclimates that produce large amounts of precipitation when the mountain rises gradually. Precipitation concentration and heavy rain are the main characteristics of the BRB. The distribution characteristics of the soil and vegetation are the same as for the climate, with gradual changes from the southeast to the northwest and obvious horizontal and vertical zonality distribution characteristics (Xie 2015). The geographic location is shown in Fig. 1.

Data sources

The BRB LUCC data were selected from Globeland30 (http://globeland30.org/) with a spatial resolution of 30 × 30 m. According to the regional characteristics and uses, the data were categorised into six types: woodland, grassland, arable land, construction land, water and other. Next, 10 land use driving factors related to climate, environment and social economy were selected. Climatic and environmental factors included elevation, slope, average temperature in the slope direction and average precipitation. Digital elevation data (DEM) were downloaded from the geospatial data cloud (https://www.gscloud.cn/). Slope and aspect were extracted by ArcGIS software in accordance with the DEM. The mean temperature and mean precipitation were downloaded from the Resource Environmental Science and Data Centre (https://www.resdc.cn/). These were spatially interpolated 1 km raster data sets of meteorological elements generated by Anuspl interpolation software for each year. Socioeconomic factors included population density, gross domestic product (GDP) and distance from a river, town and road. Population density, GDP data, river and road data were obtained from the Centre for Resources and Environmental Science and Data Research (https://www.resdc.cn/). Urban data were extracted in accordance with the LUCC data using Euclidean distance interpolation-based raster data. The technical process is shown in Fig. 2.

Research method

GeoSOS-FLUS model

The GeoSoS-FLUS model can efficiently simulate and optimise geospatial phenomena. The artificial neural network (ANN) algorithm was used to compute the applicability range and probability of each land type. Incorporating the adaptive inertia competition mechanism of roulette wheel selection, the spatial distribution of future land use was obtained under the constraints (factors).

(1) The ANN for calculating applicability probability covers three layers: input, hidden and output. It primarily includes training and prediction, and the formula is as follows:

\(\sum _{n}m\left(p,n,i\right)=1\)	(1)
\(m\left(p,n,t\right)=\sum _{j}{\omega }_{j,n}\times sigmoid\left[{net}_{j}\left(p,i\right)\right]\)	(2)
\(=\sum _{j}{\omega }_{j,n}\times \frac{1}{1+{e}^{-{net}_{j}(p,i)}}\)	(3)

Here, \(m\left(p,n,i\right)\) is the applicability probability of n type land use under set time i and grid p; \({\omega }_{j,n}\)is the weight; \(sigmoid\left( \right)\)is the associated correlation function; and \({net}_{j}(p,i)\) represents the information gained from the jth hidden layer of grid p at time i.

(2) Adaptive inertial competition mechanism

The conversion between land use will be affected by the adaptive probability, with the primary influencing factors including the competition between land type and conversion cost, neighbourhood density, inertia coefficient, etc. In the iterative process, the amount of land and demand for land were adjusted adaptively and then the inertia coefficients of different types of land use were determined. The adaptive inertia coefficient of \({Intertia}_{k}^{t-1}\) is expressed as follows:

\({Intertia}_{k}^{t-1}\left\{\begin{array}{c}{Intertia}_{k}^{t-1} \left(\left|{D}_{k}^{t-2}\right|\le \left|{D}_{k}^{t-1}\right|\right) \\ {Intertia}_{k}^{t-1}\times \frac{{D}_{k}^{t-2}}{{D}_{k}^{t-1}} \left({0>D}_{k}^{t-2}>{D}_{k}^{t-1}\right) \\ {Intertia}_{k}^{t-1}\times \frac{{D}_{k}^{t-1}}{{D}_{k}^{t-2}} \left({D}_{k}^{t-1}{<D}_{k}^{t-2}<0\right) \end{array}\right.\)

(4)

Here, \({D}_{k}^{t-1}\)and \({D}_{k}^{t-2}\)are the demand and actual grid quantity of the land k at time t-1 and t-2, respectively.

InVEST model

The InVEST model can be used to predict alterations in the quality and value of ecosystem services under various scenarios, as well as offer an objective foundation for decision-makers to balance the gains and effects of human activities. In this study, the C storage module of the InVEST model was employed to evaluate the C storage of the BRB in Gansu province. The regional C storage was estimated by combining the land use data and C density. In ecosystem C storage, there are four basic C pools. The aboveground vegetation C pool contains C from all aboveground living vegetation. The underground vegetation C pool contains C from the living roots of plants. Soil organic C pool from C in organic soil and mineral soil. Dead organic C pools include C from litter and dead plants. The formula for calculating C storage is:

\({C}_{total}=\sum _{i=1}^{n}（{C}_{i-above}{+C}_{i-below}+{C}_{i-soil}+{C}_{i-dead}）\times {A}_{i}\)

(5)

Here, \(i\) is the type of land, \({C}_{i-above}\) is the aboveground C density of the \(i\text{t}\text{h}\) species, \({C}_{i-below}\) is the underground C density of the \(i\text{t}\text{h}\) species, \({C}_{i-soil}\) is the soil C density of the \(i\text{t}\text{h}\) species, \({C}_{i-dead}\) is the dead C density of the \(i\text{t}\text{h}\) species, \({C}_{total}\) is the total C stock, \({A}_{i}\)is the area of the \(i\text{t}\text{h}\) species and n is six.

To ensure the precision of the data used in this study, C density data from Gansu Province or the same climatic zone as the study area were selected (Zhu et al. 2019). The data mainly included the analysis results of 2705 soil samples collected by Liu et al. (2012, 2013), which comprehensively considered the characteristics of soil type, parent material, topography, land use status, etc. Deng et al. (2022) used China’s arid zone data and the 2010 C density data set of the China Terrestrial Ecological System (http://www.cern.ac.cn) as a reference and selected typical soil samples to construct C density data for the Qilian Mountains (J. Deng et al. 2022). Ren et al. (2021) determined the C density of Gansu Province by modifying the national C density, mainly via reference to the relevant studies of Alam et al. ((Alam et al. 2013; Ren et al. 2021)Giardina and Ryan (2000) studied the relationship between precipitation, biomass and soil C density, while Chen et al. (2007) explored the correlation between temperature and biomass C density (G. Chen et al. 2007; Giardina and Ryan 2000; Raich and Nadelhoffer 1989)). In the present study, C density data were obtained by integrating literature and actional regional actual conditions, as shown in Table 1:

Table 1

Carbon density data.
Land use type	C_above	C_below	C_soil	C_dead
Arable land	4.56	7.45	18.4	0
woodland	40.72	13.2	46.9	6.52
Grassland	0.86	5.38	17.9	1.29
Water	0	0	0	0
Construction land	0	0	33.5	0
Other	0.63	4.95	12.2	0

Scenario establishment

Three future scenarios were set up in this study, the Natural Development scenario (ND), Ecological Protection scenario (EP), and Arable Land Protection scenario (ALP).

In the ND scenario, full consideration was given to the development needs of the BRB. Following the current land use structure and its natural evolution, the law was that the land use transformation should be the same as in the past, with a small transformation ratio between forest land and construction land, arable land and woodland. Thus, the natural development of the watershed space was modelled.

In the ALP scenario, the arable land in the whole BRB was prioritised. The transformation of arable land to other land use types was regulated to slow down the growth rate of construction land according to the relevant protection policies for arable land and the Plan of Territorial Spatial Ecological Restoration of Gansu Province (2021–2035).

In the EP scenario, the growth rate of construction land was slowed and some functions of the ecosystem gradually recovered. At the same time, the afforestation policy was strengthened and optimised. In accordance with the general plan of land, the protection of woodland, grassland and water areas was enhanced to reduce grassland degradation, woodland reduction and the trend of construction land expansion. The forest land area gradually increased, the cultivated land remained in nearly its original state and the ecosystem improved.

Temporal and spatial characteristics of C storage

As shown in Table 2, in 2000, 2010 and 2020, the C storage of the BRB was 177.39, 183.53 and 183.24 Tg respectively, an increasing trend. The cumulative C storage increased by 5.84 Tg, or 3.19%. From 2000 to 2010, the C storage increased by 6.24 Tg or 3.35%. There was a 0.3 Tg decrease in C storage from 2010 to 2020. There were changes in the C storage by different land use types in the BRB in different periods. Forest land had the largest C storage ratio, comprising about 92% of the total C stocks. The proportion of C stores decreased from arable land to grassland to construction land, while water areas and other land had very small C stores. The vegetation and soil C storage of arable land and woodland increased first and then decreased from 2000 to 2020. The vegetation, soil and total C storage of grassland decreased by 0.52, 1.57 and 2.09 Tg, respectively, while the soil C storage and total C storage of construction land increased by 0.27 Tg.

Table 2

Changes in land use types and C storage, 2000–2020.
	Vegetation C storage (Tg)				Soil C storage (Tg)			Total C storage (Tg)
	2000	2010	2020		2000	2010	2020	2000	2010	2020
Arable land	4.82	4.88	4.82		7.38	7.48	7.38	12.20	12.36	12.20
Woodland	80.87	84.91	84.72		80.12	84.12	83.94	160.99	169.03	168.66
Grassland	1.00	0.48	0.48		3.06	1.49	1.49	4.06	1.97	1.97
Water	—	—	—		—	—	—	—	—	—
Construction land	—	—	—		0.13	0.16	0.40	0.13	0.16	0.40
Other	—	—		—	—	0.01	—	0.01	0.01	0.00
Total	86.69	90.28		90.03	90.71	93.26	93.21	177.39	183.53	183.24

In terms of spatial distribution, C storage in the BRB was ‘high in the southwest and low in the northeast’, showing obvious spatial heterogeneity (Fig. 3). Taking the main trunk of the Bailong River as the boundary, the basin was divided into two parts. The high-density C storage area was primarily distributed in the west, northwest and southwest, including in Zhouqu County, Dibe County and Wen County, etc. The northwest was a concentrated area of C storage with high altitudes and high vegetation coverage, with large areas of woodland and grassland. The low-density C storage area was centred in the eastern and north-eastern parts of the basin, including in Wudu District and Tongchang District. The eastern part is highly urbanised and mainly covered by farmland, indicating that the C storage capacity of the low-density area was relatively low and vastly influenced by human activities, thereby maintaining a low level of C storage.

In terms of the changes in C storage, the increase in C storage during 2000–2010 was mainly in the northeast and east, and a small amount of increased C storage was in the northwest of Wenxian County (Fig. 4). The area with increased C storage was comparatively small. The area with decreased C storage was primarily distributed sporadically along both sides of the BRB, most obviously in Dangchang County in the northeast. In general, the C storage from 2000 to 2020 showed an increasing trend, which was distributed in the east of the BRB. Most C stocks remained relatively stable.

Impacts of land use type changes on C storage

Temporal and spatial LUCC

The area, proportion and distribution of land use types in the BRB were obtained through the land use classification results from 2000 to 2020. The quantitative analysis revealed the river basin to be primarily woodland and arable land, accounting for 74.88% and 19.48% of the total area, respectively. The construction land area was small, at only 1%. The area of arable land, woodland and other land first increased and then decreased, while the area of water and construction land increased and the area of grassland decreased.

The BRB has a long, narrow fusiform trend from the northwest to the southeast. Its terrain is high in the northwest, low in the southeast and undulating. The vegetation coverage is very dense, with woodland accounting for more than 70% of the area and mostly concentrated in the northwest, west and south. Some of the arable land and construction land was in the midlands, while some was concentrated in the north and east. In contrast to the north-western region, the western and southern regions have flat terrain that is more suitable for homes and cultivation (Fig. 5).

Impacts of land use transformation on C stocks

During 2000–2020, the area of land use in the BRB that transferred to another type was 1933.22 km² (Table 3). The area of grassland that was transferred was 945.91 km², primarily into woodland and arable land with the area that was transformed into woodland accounting for 82.31% of the total transferred area of grassland. Therefore, the vegetation C storage of grassland decreased by 3.339 Tg, the soil C storage increased by 2.313 Tg and the total C storage decreased by 5.6523 Tg. The change in grassland C storage was the largest because of its transfer to woodland. When other land types were transformed into grassland, the vegetation C storage decreased by 0.3136 Tg, soil C storage increased by 0.4594 Tg and the total C storage increased by 0.7729 Tg. As the increase in C storage was small, but the decrease was large, the total C storage of grassland decreased overall.

The conversion area of woodland and arable land was 419.11 and 545.66 km², respectively (Table 3). The area of woodland that was transformed into arable land was 303.38 km², while the area of arable land that was transformed into woodland was 449.13 km². The arable land area decreased by 3.4235 and 1.7855 Tg, respectively, basically via the transformation to woodland. The C storage of other land types that transferred to arable land increased to some extent, but the increase was less than the decrease, which resulted in an overall reduction in the C storage of arable land. The amount of mutual conversion between arable land and woodland was greater than that between other land types. The area of woodland that was converted into cultivated land was less than the area of arable land that was transformed into woodland, indicating that the total C storage of the woodland and vegetation C storage were increasing. Construction land expanded outward but the area occupied by other land types was small, so the change in C storage was also small and the changes were not obvious. The transferred areas of construction land, water areas and other land were small and the impact on the change of C stocks was small. In general, from 2000 to 2020, the overall trend in C stocks in the BRB was a slight increase, which was closely associated with the extent of land use conversion and mutual conversion.

Table 3

Land transfer and C changes in the BRB, 2000–2020.
Type of land use			Vegetation C storage		Soil C storage		Total C storage		Change in area (km²)	Change in area (%)
Transfer out		Transfer to	Increase	Decrease	Increase	Decrease	Increase	Decrease	Change in area (km²)	Change in area (%)
Arable land	Woodland		1.5730	—	—	1.8824	—	3.4554	449.13	82.31
	Grassland		0.0024	—	0.0176	—	0.0152	—	30.48
	Water		—	0.0204	0.0133	—	0.0337	—	11.08
	Construction land		0.0830	—	0.0660	—	—	0.0170	54.97
	Other		—	—	—	—	—	—	0.01
									545.66
Woodland	Arable land		—	1.0625	1.2716	—	2.3341	—	303.38	72.39
	Grassland		—	0.3175	0.4423	—	0.7598	—	92.76
	Water		—	0.0903	0.0912	—	0.1815	—	16.91
	Construction land			0.0119	0.0323	—	0.0442	—	5.99
	Other			0.0003	0.0004	—	0.0007	—	0.08
									419.12
Grassland	Arable land		—	0.0182	—	0.1330	—	0.1148	230.44	24.36
	Forst		2.3187	—	—	3.2298	—	5.5485	677.33	71.61
	Water		—	0.0239	0.0078	—	0.0316	—	12.43
	Construction land		0.0367	—	0.0160	—	——	0.0207	25.67
	Other		—	—	—	—	—		0.04
									945.91
Water	Arable land		0.0077	—		0.0050		0.0127	4.18	37.46
	Forst		0.0248	—	—	0.0250		0.0498	4.64	41.58
	Grassland		0.0014	—		0.0005		0.0019	0.75
	Construction land		0.0053	—	—	—		0.0053	1.59
									11.16
Construction	Arable land		—	0.0117	—	0.0093	0.0024	—	7.75	82.45
	Forst		0.0030	—	—	0.0081	—	0.0111	1.501	15.97
	Grassland		—	—	—	—	—	—	0.02
	Water		—	0.0004	—	—	0.0004	—	0.12
									9.40
Other	Arable land		0.0001	—	—	0.0001	—	0.0001	0.09
	Forst		0.0009	—	—	0.0011	—	0.0020	0.22
	Grassland		0.0002	—	—	—	—	0.0002	0.22
	Water		—	0.0018	0.0008	—	0.0026	—	1.45	73.23
									1.98
Sum			2.4499		3.3132		5.7631		1933.22

Spatio-temporal characteristics of LUCC and C storage under multi-scenario simulations

The GeoSOS-FLUS model was used to simulate land use under various scenarios from 2030 to 2050. First, the simulated land use status map in 2020 was compared to the actual land use in 2020 (Fig. 6). The Kappa value was 0.834 and the overall accuracy was 93.68%; thus, the simulation could feasibly and credibly be applied to predict future scenarios. Using the land use data in 2020, the LULC spatial and temporal pattern of the BRB was predicted under the ALP scenario, ED scenario and ND scenario, and the spatial changes in future C storage were obtained for each scenario using the InVEST model (Fig. 8). The regions with obvious changes in the BRB, Dibu and Dangchang counties, are amplified in Fig. 8.

Carbon storage increased in general. From 2000 to 2020, C storage increased from 177.39 to 183.24Tg. The urban development in this region is slow and remote; the land use types mainly comprise woodland, grassland and cultivated land, with a small proportion of construction land. Although the city has been in a state of expansion, compared with the strong C sequestration capacity of forest land, and the mutual conversion of land types, the overall C storage increased.

As shown in Figs. 7 and 8, under the ALP scenario, it was predicted that the C storage of the BRB in 2030 would be 183.091 Tg. This was a slight decrease f 0.147 Tg compared with 2020. In 2040 and 2050, the C storage increased slightly by 0.01 and 0.006 Tg, respectively. The change in arable land was more obvious in the surrounding areas of the Gangchang District, Wudu District and other urban areas. In terms of urbanisation and industrialisation, arable land cannot always avoid being developed and occupied. Then, when the problem of grain yield reduction results from the decrease in cultivated land, other land types that are far away from construction land need to be changed to cultivated land to achieve the required area of arable land. Therefore, the C storage under the ALP scenario kept fluctuating. Under the EP scenario, the expansion of construction land and the conversion of other land types into forest land were controlled, the protection of woodland and grassland was strengthened and the transfer to construction land and cultivated land was controlled. Between 2030 and 2050, the overall C storage increased. Compared with 2020, C storage was predicted to increase by 0.296 Tg in 2030, 0.222 Tg in 2040 and 0.160 Tg in 2050. Under the ND scenario, there was an overall increase in C storage from 2000 to 2020. The C storage was predicted to increase in 2030, but only by 0.009 Tg. In 2040, the C storage would start to decline, decreasing by 0.019 Tg, followed by an additional decrease of 0.001 Tg in 2050.

Therefore, under the EP scenario, C storage in the river basin continuously increased. In the ND scenario, after 30 years, construction development was increasingly rapid, the encroachment of construction land into cultivated land and woodland was hastened and the process of urbanisation was accelerated. This led to a gradual decrease in C storage in the BRB region. Under the ALP scenario, it was predicted that C storage in the BRB would decline. Thus, C storage only increased under the EP scenario. The large amount of C storage provided by woodland and grassland protected the C sink ability of the BRB and the EP measures obtained relatively obvious ecological benefits. In comparison to the ND scenario, the ALP scenario limited the urban expansion to some extent and also reduced the transformation of woodland and grassland to urban areas. The vegetation C stocks were protected and the total C storage tended to increase. Therefore, adopting EP measures to protect cultivated land could effectively increase C storage and reduce C storage loss.

Effects of land use conversion on C storage

One of the factors that influence changes in C stocks is the conversion of LUCC (H. Li et al. 2021) and this has also been proven to be an important way to increase C storage (Lal 2004). This study showed that changes in woodland, arable land and grassland led to an increase in C stocks, with the primary factor responsible for increasing C stocks being the transformation between woodland and arable land. Xu et al. (2016) found that terrestrial transformation changed the vegetation and the organic C of the residual parts in the ecosystem. Land use changes may upset the balance between C inflow and outflow from soils (Guo and Gifford 2002); when grassland is converted into forest land, soil C storage increases over time. The possible mechanism underpinning this is as follows. During afforestation activities, the grassland soil is disturbed by humans and soil respiration leads to C loss. Then, during the gradual development of the forest, there is an increase in soil C stock inputs, with biomass input to the soil from the forest floor in the form of apoplastic and dead wood; consequently, the soil C stock increases (L. Deng et al. 2016). For natural forest land converted to arable land, the reduced input of soil organic matter and the high temperature of the tilled soil—which accelerates the rate of decomposition of soil organic C, as well as destabilising soil organic C and increasing infiltration—decreases the soil organic matter content in forest land (L. Deng, Shangguan, et al. 2014). In contrast, when farmland is converted into woodland, the input of soil organic C in woodland increases and the stability of soil organic matter is enhanced. Forests can fix atmospheric CO₂ in the form of organic matter into plants and soil through photosynthesis (Yin and Porporato 2023), thus enhancing the carbon storage of vegetation. Deng et al. (2014) discovered that, after the conversion of arable land to woodland, there was a small loss of soil C stock at the beginning, whereafter there was a gradual recovery to a comparable level and then an increase to produce a net C gain (L. Deng, Liu, et al. 2014). Additionally, the fungal biomass of woodland is higher than that of grassland and arable land and is positively correlated with soil C concentration (Liang et al. 2015; Wiesmeier et al. 2019).

C stock changes under multi-scenario simulations

In this study, three different scenarios were simulated. Compared with the NP scenario, carbon storage in the EP scenario continued to increase. Forests and grasslands provide important C stores that safeguard their ability to sequester C in regional ecosystems. This is the result of continued afforestation and grassland restoration, resulting in continued growth in the forest area and C accumulation (Voicu et al. 2017). Ecological measures such as converting farmland to forest land can improve C storage (Lin et al. 2022; Shang et al. 2014). Zhao et al. (2019) used the CA-Markov and InVEST models to reveal that implementing positive human ecological engineering measures could improve C storage in ecosystems (Zhao et al. 2019). Accordingly, the EP and ALP measures could productively increase C storage and consolidate the C sink capacity of the BRB ecosystem. Under the AIP situation, the C storage initially decreased slightly. Wei et al. (2014) found that the transformation of woodland or grassland to arable land resulted in an initial rapid reduction in soil storage (X. Wei et al. 2014). Since the C input of litter also decreased, the C output of the cultivated land could only improve by the destruction of soil organic matter. Additionally, the long-term effective management of farm land—for example, increasing the use of organic fertilisers in topsoil—has led to a new equilibrium point for arable C stocks and a slow rate of C accumulation (Guo and Gifford 2002). According to a previous study, soil C storage only gradually begins to increase when woodland or grassland is converted to farmland and cultivated for 30 years (L. Deng et al. 2016). Therefore, under the ALP scenario, C storage will decrease to some extent and then increase gradually, which provides a reliable basis for predicting the future trend of C storage. However, ALP limited urban expansion and reduced the conversion of woodland and grassland to cities, resulting in the protection of vegetation C storage, with a slight increase in the total C storage.

Deficiency and prospects

The GeoSOS-FULS and InVEST models were used to explore the current and predicted impact of LUCC on C stocks. Uncertainty in the research results mainly came from errors in the C density data, land use data and the parameters required by the model. Additionally, the InVEST model C storage module only refers to C density, ignoring the spatial heterogeneity between the interior and the differences in carbon sequestration function caused by vegetation age (Hou et al. 2022). However, the mechanism by which the resulting changes affect C density when the area of interconversion between local classes is relatively small is still unclear (Kong et al. 2018). All these problems need to be solved sequentially based on an increased understanding of the mechanism of the influence of the changes between different species on the carbon budget. Therefore, in future studies, sufficient data should be obtained through field monitoring to verify the rationality of the C density value. Additionally, to supplement the information on spatial heterogeneity among terrestrial land use types, the impact of vegetation age structure and soil on C density would strengthen the accuracy of regional ecosystem C storage assessment.

From 2000 to 2020, the total C sequestration in the BRB increased from 177.39 to 183.24 Tg. The variation trends in vegetation C storage, soil C storage and overall C storage were roughly the same, showing the distribution characteristics of ‘high in the southwest and low in the northeast’.

During 2000–2020, there was a 1933.22 km² shift in the land use type of the BRB. The grassland outflow was larger than the transfer amount, which resulted in a decrease in the overall C storage of grassland. The conversion area of forest land into arable land was less than that of arable land into forest land, which indicated that the total C storage of forest land and vegetation C storage were increasing. The transformation amount of construction land, water area and other land was small, which approximately resulted in the C storage reaching the status of ‘balance of occupation and supplementation’.

By comparing the EP, ALP and ND scenarios it was found that, under the EP scenario, C storage continued to increase and the C sink capacity of the ecosystem was greater. Therefore, the adoption of ecological protection measures would result in relatively obvious ecological benefits.

Acknowledgements We thank Globeland30 for providing land use data. we also thank Centre for Resources and Environmental Science and Data Research for providing socio-economic and meteorological data.

Author contribution Wanli Wang: methodology, software, formal analysis, visualization, writing—original draft preparation, review and editing. Zhen Zhang: Conceptualization, validation, data curation, supervision, funding acquisition, writing—original draft preparation, review and editing. Jing Ding and Xiao Liu: project administration, funding acquisition, supervision. Heling Sun, Chao Deng and Guolong Li: visualization, writing—review. All authors read and approved the final manuscript.

Funding: This study was funded by the Natural Science Research Project of Anhui Educational Committee (Grant No.2022AH040111), National Natural Science Foundation of China (Grant Nos.42071085,41701087,52104172), Natural Science Foundation of Anhui Province (Grant No. 2108085QE207).

Data availability The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.

All authors have read, understood, and have complied as applicable with the statement on “Ethical responsibilities of Authors” as found in the Instructions for Author.

Ethical approval Not applicable.

Consent to participate Not applicable.

Consent to publication Not applicable.

Conflict of Interest The authors declare no competing interests.

Alam, S. A., Starr, M., & Clark, B. J. F. (2013). Tree biomass and soil organic carbon densities across the Sudanese woodland savannah: A regional carbon sequestration study. Journal of Arid Environments, 89, 67–76. https://doi.org/10.1016/j.jaridenv.2012.10.002
Castellano, G. R., Santos, L. A., & Menegário, A. A. (2022). Carbon Soil Storage and Technologies to Increase Soil Carbon Stocks in the South American Savanna. Sustainability, 14(9), 5571. https://doi.org/10.3390/su14095571
Chen, G., Yang, Y., Liu, L., Li, X., Zhao, Y., & Yuan, Y. (2007). Research Review on Total Below Ground Carbon Atbcation in Forest Ecosystems. Journal of Subtropical Resources and Environment.
Chen, Z., Huang, M., Zhu, D., & Altan, O. (2021). Integrating Remote Sensing and a Markov-FLUS Model to Simulate Future Land Use Changes in Hokkaido, Japan. Remote Sensing, 13(13), 2621. https://doi.org/10.3390/rs13132621
Climate Change 2007:The Physical Science Basis. (2007). World Environment, No.105(02), 13–22.
Deng, J., Ding, W., Pu, X., Lv, Y., & Wang, Y. (2022). Spatial-temporal Distribution of Carbon Storage in Qilian Mountain National Park Based on InVEST Model. Bulletin of Soil and Water Conservation. https://doi.org/10.13961/j.cnki.stbctb.20220524.001
Deng, L., Liu, G., & Shangguan, Z. (2014). Land-use conversion and changing soil carbon stocks in China’s ‘Grain-for-Green’ Program: a synthesis. Global Change Biology, 20(11), 3544–3556. https://doi.org/10.1111/gcb.12508
Deng, L., Shangguan, Z., & Sweeney, S. (2014). “Grain for Green” driven land use change and carbon sequestration on the Loess Plateau, China. Scientific Reports, 4(1), 7039. https://doi.org/10.1038/srep07039
Deng, L., Zhu, G., Tang, Z., & Shangguan, Z. (2016). Global patterns of the effects of land-use changes on soil carbon stocks. Global Ecology and Conservation, 5, 127–138. https://doi.org/10.1016/j.gecco.2015.12.004
Fang, J., Yu, G., Ren, X., Liu, G., & Zhao, X. (2015). Carbon Sequestration in China’s Terrestrial Ecosystems under Climate Change ——Progress on Ecosystem Carbon Sequestration from the CAS Strategic Priority Research Program. Bulletin of Chinese Academy of Sciences. https://doi.org/10.16418/j.issn.1000-3045.2015.06.019
Giardina, C. P., & Ryan, M. G. (2000). Evidence that decomposition rates of organic carbon in mineral soil do not vary with temperature. Nature, 404(6780), 858–861. https://doi.org/10.1038/35009076
Gomes, L. C., Faria, R. M., de Souza, E., Veloso, G. V., Schaefer, C. E. G. R., & Filho, E. I. F. (2019). Modelling and mapping soil organic carbon stocks in Brazil. Geoderma, 340, 337–350. https://doi.org/10.1016/j.geoderma.2019.01.007
Gong, J., Xu, C., Yan, L., Zhu, Y., Zhang, Y., & Jin, T. (2021). Multi-scale analysis of ecosystem services trade-offs in an ecotone in the Eastern Margin of the Qinghai-Tibetan Plateau. Journal of Mountain Science, 18(11), 2803–2819. https://doi.org/10.1007/s11629-020-6505-5
Guo, L. B., & Gifford, R. M. (2002). Soil carbon stocks and land use change: a meta analysis: SOIL CARBON STOCKS and LAND USE CHANGE. Global Change Biology, 8(4), 345–360. https://doi.org/10.1046/j.1354-1013.2002.00486.x
Hamad, R., Balzter, H., & Kolo, K. (2018). Predicting Land Use/Land Cover Changes Using a CA-Markov Model under Two Different Scenarios. Sustainability, 10(10), 3421. https://doi.org/10.3390/su10103421
He, Q., Zeng, C., Xie, P., Liu, Y., & Zhang, M. (2018). An assessment of forest biomass carbon storage and ecological compensation based on surface area: A case study of Hubei Province, China. Ecological Indicators, 90, 392–400. https://doi.org/10.1016/j.ecolind.2018.03.030
Hou, J., Chen, J., Zhang, K., Zhou, G., You, H., & Han, X. (2022). Temporal and Spatial Variation Characteristics of Carbon Storage in the Source Region of the Yellow River Based on InVEST and GeoSoS-FLUS Models and Its Response to Different Future Scenarios. Environmental Science. https://doi.org/10.13227/j.hjkx.202201267
Islam, S., Li, Y., Ma, M., Chen, A., & Ge, Z. (2021). Simulation and Prediction of the Spatial Dynamics of Land Use Changes Modelling Through CLUE-S in the Southeastern Region of Bangladesh. Journal of the Indian Society of Remote Sensing, 49(11), 2755–2777. https://doi.org/10.1007/s12524-021-01402-w
Jana, A., Jat, M. K., Saxena, A., & Choudhary, M. (2022). Prediction of land use land cover changes of a river basin using the CA-Markov model. Geocarto International, 37(26), 14127–14147. https://doi.org/10.1080/10106049.2022.2086634
Kong, J., Yang, R., Sun, Y., & Fu, Z. (2018). Effect of land use and cover change on carbon stock dynamics in a typical desert oasis. Acta Ecologica Sinica, 38(21), 7801–7812. https://doi.org/10.5846/stxb201711302156
Lal, R. (2004). Carbon Sequestration in Dryland Ecosystems. Environmental Management, 33(4), 528–544. https://doi.org/10.1007/s00267-003-9110-9
Li, H., Qu, Y., Zeng, X., Zhang, H., Cui, L., & Luo, C. (2021). Dynamic response of the vegetation carbon storage in the sanjiang plain to changes in land use/cover and climate. Heritage Science, 9(1), 134. https://doi.org/10.1186/s40494-021-00605-1
Li, W., Jia, S., He, W., Raza, S., Zamanian, K., & Zhao, X. (2022). Analysis of the consequences of land-use changes and soil types on organic carbon storage in the Tarim River Basin from 2000 to 2020. Agriculture, Ecosystems & Environment, 327, 107824. https://doi.org/10.1016/j.agee.2021.107824
Liang, C., Gutknecht, J. L. M., & Balser, T. C. (2015). Microbial lipid and amino sugar responses to long-term simulated global environmental changes in a California annual grassland. Frontiers in Microbiology, 6. https://doi.org/10.3389/fmicb.2015.00385
Lin, T., Wu, D., Yang, M., Ma, P., Liu, Y., Liu, F., & Gan, Z. (2022). Evolution and Simulation of Terrestrial Ecosystem Carbon Storage and Sustainability Assessment in Karst Areas: A Case Study of Guizhou Province. International Journal of Environmental Research and Public Health, 19(23), 16219. https://doi.org/10.3390/ijerph192316219
Liu, S. Y., Hu, N. K., Zhang, J., & Lv, Z. C. (2018). Spatiotemporal change of carbon storage in the Loess Plateau of northern Shaanxi, based on the In VEST Model. Sciences in Cold and Arid Regions, 10(03), 240–250. https://doi.org/10.3724/SP.J.1226.2018.00240
Liu, Y., Zhang, J., Zhou, D., Ma, J., Dang, R., Ma, J., & Zhu, X. (2021). Temproal and spatial variation of carbon storage in the Shule River Basin based on InVEST model. Acta Ecologica Sinica, 41(10), 4052–4065. https://doi.org/10.58466/stxb201911152452
Mamanis, G., Vrahnakis, M., Chouvardas, D., Nasiakou, S., & Kleftoyanni, V. (2021). Land Use Demands for the CLUE-S Spatiotemporal Model in an Agroforestry Perspective. Land, 10(10), 1097. https://doi.org/10.3390/land10101097
Mattsson, E., Ostwald, M., Wallin, G., & Nissanka, S. P. (2016). Heterogeneity and assessment uncertainties in forest characteristics and biomass carbon stocks: Important considerations for climate mitigation policies. Land Use Policy, 59, 84–94. https://doi.org/10.1016/j.landusepol.2016.08.026
Meeussen, C., Govaert, S., Vanneste, T., Haesen, S., Van Meerbeek, K., Bollmann, K., et al. (2021). Drivers of carbon stocks in forest edges across Europe. Science of The Total Environment, 759, 143497. https://doi.org/10.1016/j.scitotenv.2020.143497
Mumcu Kucuker, D. (2020). Spatiotemporal changes of carbon storage in forest carbon pools of Western Turkey: 1972–2016. Environmental Monitoring and Assessment, 192(8), 555. https://doi.org/10.1007/s10661-020-08431-x
Raich, J. W., & Nadelhoffer, K. J. (1989). Belowground Carbon Allocation in Forest Ecosystems: Global Trends. Ecology, 70(5), 1346–1354. https://doi.org/10.2307/1938194
Remy, E., Wuyts, K., Boeckx, P., Ginzburg, S., Gundersen, P., Demey, A., et al. (2016). Strong gradients in nitrogen and carbon stocks at temperate forest edges. Forest Ecology and Management, 376, 45–58. https://doi.org/10.1016/j.foreco.2016.05.040
Ren, X., Pei, T., Chen, Y., Xie, B., & Chen, D. (2021). Impact of land use change on carbon storage in Gansu Province based on carbon density correction. Ecological Science. https://doi.org/10.14108/j.cnki.1008-8873.2021.04.008
Sahu, B., Ghosh, A. K., & Seema. (2021). Deterministic and geostatistical models for predicting soil organic carbon in a 60 ha farm on Inceptisol in Varanasi, India. Geoderma Regional, 26, e00413. https://doi.org/10.1016/j.geodrs.2021.e00413
Shang, Z.-H., Cao, J.-J., Guo, R.-Y., Long, R.-J., & Deng, B. (2014). The response of soil organic carbon and nitrogen 10years after returning cultivated alpine steppe to grassland by abandonment or reseeding. CATENA, 119, 28–35. https://doi.org/10.1016/j.catena.2014.03.006
Tian, L., Tao, Y., Fu, W., Li, T., Ren, F., & Li, M. (2022). Dynamic Simulation of Land Use/Cover Change and Assessment of Forest Ecosystem Carbon Storage under Climate Change Scenarios in Guangdong Province, China. Remote Sensing, 14(10), 2330. https://doi.org/10.3390/rs14102330
Vizcaíno-Bravo, Q., Williams-Linera, G., & Asbjornsen, H. (2020). Biodiversity and carbon storage are correlated along a land use intensity gradient in a tropical montane forest watershed, Mexico. Basic and Applied Ecology, 44, 24–34. https://doi.org/10.1016/j.baae.2019.12.004
Voicu, M. F., Shaw, C., Kurz, W. A., Huffman, T., Liu, J., & Fellows, M. (2017). Carbon dynamics on agricultural land reverting to woody land in Ontario, Canada. Journal of Environmental Management, 193, 318–325. https://doi.org/10.1016/j.jenvman.2017.02.019
Wei, P., Chen, S., Wu, M., Jia, Y., Xu, H., & Liu, D. (2021). Increased Ecosystem Carbon Storage between 2001 and 2019 in the Northeastern Margin of the Qinghai-Tibet Plateau. Remote Sensing, 13(19), 3986. https://doi.org/10.3390/rs13193986
Wei, X., Shao, M., Gale, W., & Li, L. (2014). Global pattern of soil carbon losses due to the conversion of forests to agricultural land. Scientific Reports, 4(1), 4062. https://doi.org/10.1038/srep04062
Wiesmeier, M., Urbanski, L., Hobley, E., Lang, B., von Lützow, M., Marin-Spiotta, E., et al. (2019). Soil organic carbon storage as a key function of soils - A review of drivers and indicators at various scales. Geoderma, 333, 149–162. https://doi.org/10.1016/j.geoderma.2018.07.026
Xie, Y. (2015). Spatiotemporal Change of Ecosystem Services Based on InVEST Model in the Bailong River Watershed, Gansu. Dissertation, LanZhou University. Retrieved from https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CDFD&dbname=CDFDLAST2019&filename=1016700033.nh&v=
Yan, Y., Wang, J., Tian, D., Luo, Y., Xue, X., Peng, F., et al. (2022). Sustained increases in soil respiration accompany increased carbon input under long-term warming across global grasslands. Geoderma, 428, 116157. https://doi.org/10.1016/j.geoderma.2022.116157
Yang, J., Xie, B., & Zhang, D. (2021). Spatio-temporal evolution of carbon stocks in the Yellow River Basin based on InVEST and CA-Markov models. Chinese Journal of Eco-Agriculture. https://doi.org/10.13930/j.cnki.cjea.200746
Yin, J., & Porporato, A. (2023). Global Self-Similar Scaling of Terrestrial Carbon With Aridity. Geophysical Research Letters, 50(3). https://doi.org/10.1029/2022GL101040
Zhang, L. (2016). Ecosystem Services Assessment and its Driving Forces in the Bailongjiang Watershed of Gansu. Dissertation, LanZhou University. Retrieved from https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD&dbname=CMFD201602&filename=1016723483.nh&v=
Zhao, M., He, Z., Du, J., Chen, L., Lin, P., & Fang, S. (2019). Assessing the effects of ecological engineering on carbon storage by linking the CA-Markov and InVEST models. Ecological Indicators, 98, 29–38. https://doi.org/10.1016/j.ecolind.2018.10.052
Zhu, W., Zhang, J., Cui, Y., Zheng, H., & Zhu, L. (2019). Assessment of territorial ecosystem carbon storage based on land use change scenario: A case study in Qihe River Basin. Acta Geographica Sinica, 74(03), 446–459. https://doi.org/10.11821/dlxb201903004

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Spatio-temporal variation and future multi-scenario simulation of carbon storage in Bailong River Basin using GeoSOS-FLUS and InVEST models

Status:

Version 1

Abstract

Figures

Introduction

Study overview and methodology

Research results

Discussion

Conclusion

Declarations

References

Additional Declarations

Status:

Version 1