The impact of wind energy on plant biomass production in China

Abstract

Global wind power expansion raises concerns about its potential impact on plant biomass production (PBP). Using a high-dimensional fixed effects model, this study reveals significant PBP reduction due to wind farm construction based on 2404 wind farms, 108,361 wind turbines, and 7,904,352 PBP observations during 2000–2022 in China. Within a 1–10 km buffer, the normalized differential vegetation and enhanced vegetation indices decrease from 0.0097 to 0.0045 and 0.0075 to 0.0028, respectively. Similarly, absorbed photosynthetically active radiation and gross primary productivity decline from 0.0094 to 0.0034% and 0.0003–0.0002 g*C/m² within a 1–7 km buffer. Adverse effects last over three years, magnified in summer and autumn, and are more pronounced at lower altitudes and in plains. Forest carbon sinks decrease by 12,034 tons within a 0–20 km radius, causing an average economic loss of $1.81 million per wind farm. Our findings underscore the balanced mitigation strategies for renewable energy transition when transiting from fossil fuels.

Full text

1
Vol.:(0123456789)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports
The impact of wind energy on plant
biomass production in China
Li Gao
1,6, Qingyang Wu
2,6, Jixiang Qiu
1, Yingdan Mei
3*, Yiran Yao
1, Lina Meng
4 &
Pengfei Liu
5
Global wind power expansion raises concerns about its potential impact on plant biomass production
(PBP). Using a high-dimensional xed eects model, this study reveals signicant PBP reduction
due to wind farm construction based on 2404 wind farms, 108,361 wind turbines, and 7,904,352
PBP observations during 2000–2022 in China. Within a 1–10 km buer, the normalized dierential
vegetation and enhanced vegetation indices decrease from 0.0097 to 0.0045 and 0.0075 to 0.0028,
respectively. Similarly, absorbed photosynthetically active radiation and gross primary productivity
decline from 0.0094 to 0.0034% and 0.0003–0.0002 g*C/m2 within a 1–7 km buer. Adverse eects
last over three years, magnied in summer and autumn, and are more pronounced at lower altitudes
and in plains. Forest carbon sinks decrease by 12,034 tons within a 0–20 km radius, causing an average
economic loss of $1.81 million per wind farm. Our ndings underscore the balanced mitigation
strategies for renewable energy transition when transiting from fossil fuels.
Environmental degradations and extreme weather driven by climate change are signicant threats to biodiversity
and ecosystems1. Fossil fuel consumption has been a major contributor to this problem, leading to various forms
of pollution and greenhouse gas emissions. e Paris Agreement encourages legally binding international trea-
ties on climate change and sustainable development by enhancing the implementation of nationally determined
contributions. e Agreement also emphasizes the critical role of renewable energy in these eorts2.
Traditional fossil fuel-based power generation, especially coal-red power plants, has signicantly aected
the health of vegetation. e emissions of sulfur oxides and nitrogen oxides from plants not only lead to acid rain
but also lower the pH value of the atmosphere, subsequently damaging soil structure and biological activity3.
Such acidic conditions inhibit the uptake of vital nutrients in many plants, including calcium, magnesium, and
potassium, hindering their growth4. Moreover, heavy metals like mercury, cadmium, and lead, released from
coal-burning, accumulate in soils, could enter the food chain and pose threats to higher-order organisms5. In
addition to these direct impacts, emissions of particulates and other gases may alter local microclimates, aecting
temperature and humidity6. e particulate matter from coal combustion can also impede sunlight penetration,
resulting in reduced light intensity or shis in light quality, which, in turn, aects photosynthesis and other
physiological processes in certain plants7. Owing to these multifaceted factors, regions under long-term inu-
ence of coal-red power plants might experience a decline in biodiversity. Such pollution could render certain
plant species less viable, leading to shis in ecosystem structure and functionality.
Wind energy is one of the fastest-growing energy sources, playing a vital role in global eorts to mitigate
climate change. Wind energy is oen favored over photovoltaic solar energy for its superior grid-balancing
capabilities8,9. However, wind farms might threaten carbon sink and plant biomass production (PBP), a vital
component of carbon storage and uptake in ecosystems10–12. Wind farm construction may also aect biodiversity
by reshaping of trophic cascades13,14, introducing new species into given ecological systems15,16, altering species
migration patterns17, killing birds through collision with turbines18, habitat degradation or provision18–22, chemi-
cal element deposition23, changing local microclimates24–26, and others. Even though these negative impacts
may not deter the shi from fossil fuel reliance to cleaner energy sources, policymakers and institutions oen
prioritize renewable energy transition without considering potential ecological impacts such as PBP conserva-
tion, leading to unsustainable resource use and increasing biodiversity destruction.
OPEN
1School of Economics and Management, China University of Petroleum Beijing, Beijing 102249, People’s
Republic of China. 2Fielding School of Public Health, University of California, Los Angeles, Los Angeles,
CA 90095, USA. 3School of Applied Economics, Renmin University of China, Beijing 100872, People’s Republic
of China. 4School of Economics and The Wang Yanan Institute for Studies in Economics, Xiamen University,
Xiamen 361005, Fujian, People’s Republic of China. 5Department of Environmental and Natural Resources
Economics, University of Rhode Island, Kingston, RI 02881, USA. 6These authors contributed equally: Li Gao and
Qingyang Wu. *email: meiyingdan@ruc.edu.cn
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2
Vol:.(1234567890)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
Plants play a critical role in ecosystem functioning and global biodiversity, and are highly vulnerable to dam-
ages caused by climate change1,28. Climate change, largely accelerated by fossil fuel use, is making them increas-
ingly vulnerable. Wind farms also have mixed eects on ecosystems, they can alter local surface temperature,
humidity, and precipitation19,22,29. ey may extend regional vegetation growing seasons and increase grassland
productivity27,30, and turbine bases may provide favorable habitats for some terrestrial species31. Previous studies
found the spatial impact of wind farms on vegetation ecosystem may extend for 10km beyond the base of the
wind turbines and the tower32. Plants in developing countries are particularly vulnerable33,34. Existing literature
focuses on the eects of wind farm construction on animal diversity, and the potential impacts on vegetation
dynamics are understudied. erefore, a nuanced understanding of the ecological implications of both renew-
able and non-renewable energy is vital.
In the context of the pressing need to transition away from fossil fuels, China has the largest installed wind
energy capacity in the world, following rapid growth in new wind facilities since the rst wind farm in 198635,36.
As of the end of 2020, the country had, cumulatively, 281 gigawatt (GW) installed wind power capacity, having
added 71.6 GW of new capacity that same year; in contrast the United States had installed capacity of 118 GW,
with 14 GW of new capacity added in 202037. China also operates almost half of the world’s installed oshore
wind power, amounting to 170 GW as of 20218,38,39. is reliance on wind power is likely to grow, as the govern-
ment announced ambitious carbon emissions goals at the 75th session of the United Nations General Assembly
in 2020, pledging to peak before 2030 and achieve carbon neutrality before 2060 (UNGA 75). Leaders further
announced the goal of a non-fossil energy share of 25% in primary energy consumption by 2030 and 80% by 2060.
China’s 14th 5-Year Plan (2021–2025) aims to reduce the number of wind farm units and investment costs by
constructing wind farms with larger capacity turbines, swept areas, and higher hub heights and rotor diameters
to maintain low wind power prices. China is one of the wealthiest countries in PBP globally and a signatory to
the UN Convention on Biological Diversity40. While the emphasis on wind power is in line with global sustain-
ability goals, it’s crucial that these eorts are executed with care for local ecosystems. Balancing renewable energy
production with biodiversity conservation remains a challenge and necessary issue to address.
Previous studies on the impact of wind farms on vegetation dynamics have relied heavily on numerical simu-
lation models due to data limitations. Such studies are oen restricted to single geological types19,20, localized
areas41–43, and/or short observation intervals42. is study uses nationwide data on the installation of 2,404 wind
farms (Fig.1) and 108,361 wind turbines in China and monthly remote sensing data on PBP indicators matched
with 12 distance buer zones for each wind farm from January 2000 to October 2022, totaling 7,904,352month-
by-buer-zone samples (Fig.2). We assessed the inuence of China’s rapidly expanding wind farm initiative
on vegetation growth while estimating the heterogeneous eects on PBP and carbon sequestration, taking into
account controls for natural, environmental, and anthropogenic factors. We examine ten PBP indicators: the
normalized dierential vegetation index (NDVI), enhanced vegetation index (EVI), the fraction of absorbed
photosynthetically active radiation (FPAR), leaf area index (LAI), gross primary productivity (GPP), net pho-
tosynthesis (NP), net primary productivity (NPP), percentage of tree cover (PTC), percentage of non-tree veg-
etation cover (PNTV), and percentage of non-vegetation cover (PNV). We also assess the impact of wind farm
construction on forest carbon sinks and economic value loss. Using the accumulation volume expansion method,
we estimate the total carbon sink of forest trees based on the Ninth National Forest Inventory of China. We then
calculate the forest carbon sink loss within the 20km range of wind farms and employ the carbon tax method,
utilizing the widely used Swedish carbon tax price of 150 USD/t of carbon, to evaluate the economic value loss of
carbon sinks. Our ndings highlight that wind power development has had profound ecological consequences.
is paper makes three primary contributions. Firstly, it provides a systematic evaluation of the externalities
of wind turbines based on the wind power construction dataset and geographic information on PBP. e wind
power construction China serves as a natural experiment to evaluate potential negative externalities of wind
power on PBP within a 10km buer zone. Secondly, we evaluate the temporal dynamic eects of wind turbine
construction and captures their heterogeneous eects on season, elevation, land types, and other characteristics.
irdly, we discuss the potential negative contributions of increasing wind power to low-carbon transformation.
Specically, we estimate the total carbon sink and economic value loss of forest trees from wind turbine construc-
tion using the cumulative volume expansion and carbon tax method, respectively, which provides a reference
point for policymakers to consider the sustainability of vegetation protection and the ecological environment
and achieve low-carbon transformation goals.
Results
Changes in PBP due to wind farm construction
Wind farms could impact 0.08% of China’s terrestrial land area, or approximately 755,216 km2 if the impacts
extend 10km from each turbine, which represents an enormous spatial footprint that regional biodiversity threat
maps and renewable energy plans fail to recognize. Figure3 and Supplementary Table4 present the ndings on
the impact of wind farms on PBP, measured by a range of indicators across varying distance buers from 1 to
20km. Estimations from Eq.(1) indicate the relationship between wind farms and PBP is generally U-shaped.
e NDVI within 1km of the wind farms exhibits a signicant decrease of 0.0097 (P < 0.01; 95% CI − 0.0161
to − 0.0033), followed by declines of 0.0128 (P < 0.001; 95% CI − 0.0193 to − 0.0063) and 0.0116 (P < 0.001; 95%
CI − 0.0180 to − 0.0052) when the distances are extended to 2km and 3km, respectively. e negative eect
gradually diminishes and reaches − 0.0089 (P < 0.01; 95% CI − 0.0147 to − 0.0030) at around 4km and is gone
by 15km. e weak U-shaped relationship remains consistent across the set of outcome variables. e negative
impact of wind farms on EVI is signicant within 1km to 8km, with a peak at 2km and a maximum decrease
of − 0.0088 (P < 0.001; 95% CI − 0.0128 to − 0.0047). Similarly, the impact on the FPAR is signicant within 1km
to 7km, peaking at 1km, with a maximum decrease of − 0.0094 (P < 0.01; 95% CI − 0.0147 to − 0.0040) of a
Content courtesy of Springer Nature, terms of use apply. Rights reserved

3
Vol.:(0123456789)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
percentage point. e negative impact on the LAI is signicant within a range of 1km to 3km, peaking at 1km,
with a maximum decrease of − 0.042 (P < 0.01; 95% CI − 0.0719 to − 0.0121). Finally, the negative impact on GPP
is signicant within a range of 1km to 7km, peaking at 3km, with a maximum decrease of − 0.0004kg*C/m2
(P < 0.01; 95% CI − 0.0006 to − 0.0001). Certain PBP indicators show less consistent and pronounced variation
aer wind farm construction, including NPP, NP, and PTC.
Dynamic eects of wind farms
Figure4a and Supplementary Table5 display the changes in PBP before and aer the construction of wind farms
by plotting the event-study coecients estimated from Eq.(2) in the methods section. Figure4a indicates that
the PBP changes within 6–12months before the wind farm begins operations, likely due to related constructions
such as preparing the site, installing infrastructure such as roads and transmission lines, and erecting the wind
turbines. PBP decreases aer the installation, consistent with expectations. In the rst year aer the construction,
NDVI declined by approximately 0.0105 for proximate wind farms. ese results remain consistent outside the
3-year evaluation period.
Figure4 indicates that the construction and operation of wind turbines may have long-term, increasing
impacts on plant biodiversity. A potential cause is the impact of human activities related to road construction,
building construction, and tourism development44. Turbines may also aect the distribution and ecological niche
of plants. e noise and vibration turbines generate may damage vegetation and soil. Changes in air ow caused
by the rotation of wind turbines may aect the climate and precipitation patterns of nearby areas. In addition to
noise, light pollution and landscape alterations may negatively aect plants over time.
Heterogeneous eects by season
Figure4b and Supplementary Table6 reveal that the negative impact of wind farms on plant communities pre-
dominantly occurs during the summer (June–August) and autumn (September–November) months, compared
to the baseline. In these periods, the establishment of wind turbines reduces the FPAR component of vegetation
by an average of 0.9896% (P < 0.05) and 0.7190% (P < 0.01) at 4km distance, respectively. In contrast, the eect
Figure1. Distribution of wind farms in China. e dots (light to dark) represent wind farms installed in
dierent years. e brown curve indicates the city’s administrative boundary. e gure depicts a total of 2,404
wind farms from the original dataset, with the earliest installation date in 1994 and the latest in 2021. Figure
produced using ArcGIS Pro 3.2 (https:// www. esri. com/ en- us/ arcgis/ produ cts/ arcgis- pro/ overv iew).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

4
Vol:.(1234567890)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
is only 0.4457% (P < 0.10) and 0.3429% (P < 0.10) during the spring (March–May) and winter (December–Feb-
ruary) months. Similarly, the seasonal response of GPP to wind turbine construction is statistically signicant
only in the summer and autumn months, whereas the seasonal response of NP is signicant only in autumn.
Existing literature indicates that the construction of wind farms alters the growth environment of plant com-
munities by aecting factors such as light, temperature, humidity, and soil conditions. e rotation of wind tur-
bines generates strong winds, resulting in leaf damage and increased evaporation. Additionally, the establishment
of wind farms can disrupt the surrounding water cycle, potentially leading to soil aridity or excessive moisture.
e most signicant impacts of changes in the water cycle and light conditions occur during the summer and
autumn seasons, as plants have the highest demand for water and light in these seasons and thus the weakest
adaptability to these eects45–47. In comparison, the impact during the spring and winter seasons is relatively
minor due to lower plant growth and reproductive demands for water and light, and a reduced inuence of
Figure2. e illustration of eld design of empirical evaluation for wind farms. Notes: Examples of wind
turbine matrixes in wind farms (a) and satellite imagery of individual wind turbines (b) are provided in images
(a, b). e red square indicates the area of inference. Images (a, b) were obtained from Google Earth (version
7.1.5.1557). e experimental design pattern is displayed in the vertical view (c, d). e experimental annuli
were divided along the radius to the center of a wind turbine using dashed circles and curves. e method used
to calculate changes in PBP around wind farms is illustrated in images c and d. e mean values of several PBP
indicators within a radius of 1km, 1–2km, …, 9–10km, 10–15km, and 15–20km from the center of the wind
farm were computed for each wind farm (c). Specically, the mean value of the PBP indicators was computed
in an area consisting of 1km and 2km from the center of the wind farm, excluding the area within 1km (the
shaded region) to illustrate the approach (d). is approach enabled the comparison of the marginal variation of
indicator values within each distance bin. Figure produced using ArcGIS Pro 3.2 (https:// www. esri. com/ en- us/
arcgis/ produ cts/ arcgis- pro/ overv iew).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

5
Vol.:(0123456789)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
wind farms during these periods24,48. is result has signicant implications for the planning and management
of wind farm construction to minimize the negative externality to the surrounding natural environment19,30,49.
Heterogeneous eects by elevation
Elevation is a crucial factor inuencing the impact of wind farms on vegetation50, as Fig.4c and Supplementary
Table7 suggest. Compared to the reference baseline regression, the construction of wind farms impacts plant
communities at elevations lower than 500m, but not at higher elevations. For plant communities below 500m,
Figure3. Average impacts of wind farm on PBP by distance. Notes: e graphs depict regression results for 10
PBP indicators, including the normalized dierence vegetation index (NDVI), enhanced vegetation index (EVI),
fraction of photosynthetically active radiation (FPAR), leaf area index (LAI, %), gross primary product (GPP,
gC/m2), net photosynthesis (NP, gC/m2), net primary productivity (NPP, kg*C/m2), percentage of tree cover
(PTC, %), percentage of non-tree vegetation (PNTV, %), and percentage of non-vegetation (PNV, %). e red
points represent point estimators, and the green spikes show the 95% condence intervals.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

6
Vol:.(1234567890)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
the establishment of wind turbines reduces the FPAR component of vegetation by an average of 1.2792% (P < 0.01)
at a distance of 4km. is eect is only 0.1350% and 0.4287% for plant communities at elevations between 500
and 1,500m and elevations above 1,500m, respectively, and is statistically insignicant. Similarly, the installation
of wind turbines only inuences the GPP and NP of plant communities at altitudes lower than 500m, while this
eect on communities above 500m is not signicant.
At lower elevations, plant communities are likely to be closer to wind turbines and more exposed to strong
winds. e rotation of wind turbine blades may generate more disturbances to vegetation, potentially causing
tree swaying or leaf damage, subsequently adversely aecting vegetation growth51,52. Moreover, at lower eleva-
tions, higher relative humidity, fog, and clouds are more common, which increase moisture and humidity on
the plant leaf surfaces, thus elevating the risk of plants swaying in high-speed winds19,53. In addition to the
Content courtesy of Springer Nature, terms of use apply. Rights reserved

7
Vol.:(0123456789)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
more dispersed distribution of plant communities at higher elevations, harsher climatic conditions and dierent
vegetation types may result in more resilient ecosystems and a more robust response to external disturbances,
including wind farms.
Heterogeneous eects by land types
In our study, land type strongly inuenced the impact of wind turbine installation on vegetation FPAR compo-
nents within a few kilometers, as Fig.4d and Supplementary Table8 suggest. e construction of wind farms
had a stable impact on plant communities in plain land types but not plateaus, hills, low-relief mountains, or
medium-relief mountains. e establishment of wind turbines led to an average reduction of 1.4859% (P < 0.01)
in vegetation PAR components at 4km, and persisted up to 20km for plain land type. For plateau, the impact
of wind turbines was only present within areas smaller than 3km, with a coecient size of 0.6969% (P < 0.05),
while the corresponding response in plant communities of plain land type at the same distance was 1.4034%
(P < 0.01). e eects of wind turbine installation on gross primary production and net photosynthesis of plant
communities were only signicant in plain land type, with eect sizes of 0.8392 gC/m2 and 0.4822 gC/m2 at 4km,
respectively. No signicant impact was observed in plant communities of other land types.
Dierences in species composition and adaptability of plant communities in dierent land types may explain
the diering impact of wind farms. Plains oen have large areas of farmland and grassland, with a relatively
simple vegetation structure, making the plant communities more vulnerable to external disturbances and dam-
age. By contrast, plateaus, hills, low and medium-relief mountains have a more complex vegetation structure,
including a higher diversity of shrubs, herbaceous plants, and trees that are better adapted to the eects of wind
farms19,54. Furthermore, plain land types generally have at terrain and relatively deeper soils, which facilitate the
interception of water and nutrients by equipment and structures associated with wind farms, potentially caus-
ing damage to the original vegetation19,55,56. Plateaus, hills, low-relief mountains, and medium-relief mountains
typically exhibit greater elevation dierences and slopes, and the plant communities within these areas tend to
be more uniform, making it more dicult for wind farm equipment and structures to negatively impact the
vegetation57,58. ese factors contribute to the increased vulnerability of plant communities in plain land types to
the adverse eects of wind farm construction, while plant communities in plateaus, hills, low and medium-relief
mountains are more resilient to such impacts59–61.
Discussion
Our study focuses on the underrepresented implications of wind energy on vegetation dynamics. Results reveal
that wind farm construction has a signicantly negative eect on PBP, and the extent of these impacts varies
across indicators. On average, wind farm construction leads to a decrease in NDVI and EVI within a range of
1–10km, from 0.0097 to 0.0045 and from 0.0075 to 0.0028, respectively. It also leads to a decrease in FPAR
and GPP within a range of 1–7km, from 0.0094% to 0.0034% and from 0.0003 to 0.0002g*C/m2, respectively.
e study also shows that wind farm eects follow a weak U-shaped relationship with several indicators, with
the eects on NDVI and EVI reaching their maximum at 2km. On average, wind farms result in a decrease in
non-tree vegetation cover, and an increase in the proportion of non-vegetation cover, with no signicant impact
on tree cover, suggesting that vegetation with lower height, such as grassland and shrubs, is more vulnerable.
Our dynamic analysis indicates that certain diversity indicators exhibit a signicant decrease within six
months before the wind farm operation, suggesting that the construction process itself has some detrimental
Figure4. Heterogeneous eects of wind farm on PBP as a function of distance. Note: Vertical axis represents
each distance range in kilometers and the horizontal axis represents each index of PBP. (a) Depicts the
dynamic eects of wind farm on PBP by distance. For each indicator, estimates of ve temporal phases are
reported by distance: “ − 6 to 0months” or “ − 1year,” indicating half a year or one year prior to full wind farm
installation; “1st,” “2nd,” “3rd,” and “ > 3year,” representing one, two, three years, or more than three years
post full installation of the wind farm. (b) Illustrates the seasonal heterogeneous eects of wind farm on PBP.
For each indicator, estimates for the “post” phase in subsamples of the four seasons are reported by distance:
“March to May” (Spring), “June to August” (Summer), “September to November” (Autumn), and “December to
February” (Winter). (c) Represents the heterogeneous eects of wind farm elevation on PBP. For each indicator,
estimates for the “post” phase in subsamples of three elevation (EL) ranges are reported by distance: “EL less
than 500m,” “EL between 500 and 1,500m,” and “EL greater than 1500m.” (d) Denotes the heterogeneous
eects of wind farm land type on PBP. For each indicator, estimates for the “post” phase in subsamples of ve
dierent land types are reported by distance: “Plain,” “Terrace,” "Hill," "Low relief mountain," and "Intermediate
relief mountain." Dierent bubbles represent normalized [0,1] values of estimated coecients based on dierent
plant indicators. e coecients of all heterogeneity groups and 12 zones for each indicator are normalized as
one group (and separated by dashed lines), so the size of the coecients within each group is comparable, while
bubbles between indicators are not comparable, and negative coecients will inevitably have smaller bubbles
than positive ones. Bubble color transitions from dark to light signify varying levels of statistical signicance;
darkest representing signicant at 1%, followed by 5%, and 10%; blank indicates non-signicance at the 10%
statistical level. is gure reveals the heterogeneity of the wind farm’s interstitial eects on each indicator,
the dierences in distance, and the relative size variations of the coecients. All regressions include the same
explanatory variables, including NDVI, EVI, FPAR, LAI, GPP, NP, NPP, PTC, PNTV, and PNV. Dierent
shades from black to light represent dierent levels of statistical signicance, with black representing a 1%
level of statistical signicance. e absolute values of the specic eects can be found in the corresponding
Supplementary tables.
◂
Content courtesy of Springer Nature, terms of use apply. Rights reserved

8
Vol:.(1234567890)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
eects on vegetation. Furthermore, the negative impacts persist and even increase for more than three years
aer the wind farm become operational. Heterogeneity analysis reveals that the eects of wind farms on plant
communities varied according to season, altitude, and land type. Negative impacts on vegetation photosynthesis
and production are more pronounced during the summer and autumn months, at lower altitudes, and in plain
land types. Dierent types of plants may exhibit distinct responses under these specic conditions. Regarding
the underlying reasons, rstly, summer and autumn oen experience higher temperatures and lower precipita-
tion levels. Plants during these periods might be in crucial growth stages, such as owering or fruit matura-
tion, rendering them more sensitive to environmental factors. Furthermore, summer and autumn are likely
peak operational times for wind farms, as increased electricity generation is required during high-temperature
weather. Elevated wind turbine activity could lead to more pronounced environmental disturbances. Secondly,
lower-altitude regions typically feature warm, arid climates and fertile soils, though they might also be more
susceptible to erosion and nutrient loss. Moreover, the plant communities in these areas might be more prone
to disturbances due to their adaptation to relatively stable environmental conditions. Lastly, ecosystems in plain
regions might be comparatively simple, lacking sucient biodiversity to withstand environmental disruptions.
ey oen face greater susceptibility to anthropogenic land use changes, such as agriculture and urbanization,
which could interact synergistically with wind farm development. ese ndings highlight the importance of
considering the heterogeneous eects of wind farm construction on the surrounding natural environment in
planning and management eorts to minimize negative impacts.
While the longstanding damage caused by coal-red power plants on vegetation has been extensive and
alarming, from an economic perspective, the research ndings are also valuable to assess the damage to vegeta-
tion productivity and corresponding environmental and economic costs from the construction of wind farms
in China. According to the United Nations Framework Convention on Climate Change, carbon sinks reduce the
amount of carbon dioxide in the atmosphere by absorbing and storing carbon dioxide through photosynthesis in
forest trees62–65. Forest carbon sequestration is more sustainable and stable than other methods of carbon reduc-
tion, lasting for hundreds of years, and can eectively mitigate global climate change66. However, Holst found
that the growing environment of forest trees can inuence the capacity of carbon sequestration, and the external
eects, such as forest management, forest land use, and other human interference can alter the quality and areas
of forest trees, resulting in variation in forest carbon sink67. us, we calculate the change of forest areas due to
the construction of wind farms with the estimated coecients of the PTC and then compute the change of forest
carbon sink to assess the economic value of lost forest carbon sink due to wind farms68.
Using the accumulation volume expansion method, we estimate the total carbon sink of forest trees. e
procedure (shown in Eq.(1) below) involves calculating forest biomass carbon sequestration based on the volume
expansion coecient, converting the biomass into biomass dry weight using volume density, and determining
carbon sequestration by applying the rate of carbon content. e carbon sequestration amounts of understory
plants and forest land are then computed using the respective carbon conversion coecients, resulting in the
total carbon sink of forest trees.
where
CF
is the total carbon sink of forest trees;
S
is the forest area in the buer zone;
C
is the forest carbon
density;
α
is the understory plants carbon conversion coecient (taking the value of 0.195);
β
is the forest land
carbon conversion coecient (taking the value of 1.244);
V
is the per unit area of forest volume (taking the value
of 79.66m3/hm2);
δ
is the volume expansion coecient (taking the value of 1.90);
ρ
is the volume density (tak-
ing the value of 0.5);
γ
is the rate of carbon content (taking the value of 0.5). is method utilizes the per unit
area of forest volume data from the Ninth National Forestry Survey, which reports a standard value of 79.66 m3
per hectare for China, along with conversion coecients specied by the Intergovernmental Panel on Climate
Change (IPCC)69. We obtain the amount of forest carbon sink loss in each buer zone within the 20km range
of the wind farm by calculating the amount of change in the percentage of tree cover, multiplied by unit area
accumulation in China. Furthermore, we employ the carbon tax method to calculate the economic value loss of
carbon sinks based on Swedish carbon tax price of 150 USD/t.
According to Supplementary Table4, within the study area of 0–20km from the wind farm, the average for-
est carbon sink decreases by 12,034.21 tons aer construction, and the average economic loss per wind farm of
carbon sink reaches $1.81 million. e largest decrease in tree cover is observed in the 9–10km buer, with an
average decrease of 0.27%, and the average carbon sink loss reaches $225,715. However, the average tree cover
increases by 0.05% in the 3–4km buer, and the average carbon sink economic value increases by $14,027.
Overall, except for the average carbon sink increase in the 2–4km buer, other buer zones are showing dierent
degrees of carbon sink loss. According to Supplementary Tables5, 6, 7 and 8, the loss of forest carbon sink in one
year prior to the construction of wind farms is the smallest, with an average carbon sink reduction of 2109.52
tons and an average economic loss of $316,427.44. e loss is the largest aer three years of construction, with
an average carbon sink reduction of 18,158.28 tons, and an average economic loss of $2.72 million. e total
reduction of carbon exceeds 10,000 tons per year, with the largest reduction of 14,284.93 tons and economic
loss of $2.24 million in spring, and the smallest reduction in autumn. Wind farms at greater elevation areas have
the largest impact on the carbon sink, with an increase of 11,539.22 tons and an average economic gain of $1.73
million, while wind farms at intermediate elevation areas have the smallest impact on the carbon sink with a
reduction of 9,622.30 tons and an average economic loss of $1.44 million. Wind farms in plains, hills, and low
relief mountains are leading to a loss of forest carbon sinks while wind farms in plateaus and medium-relief
mountains are not. e greatest loss is in the hilly areas, amounting to 52,657.06 tons and an average economic
loss of $7.90 million.
(1)
C
F=
(S
×
C)
+
α(S
×
C)
+
β(S
×
C)
C
=
V
×
δ
×
ρ
×
γ
Content courtesy of Springer Nature, terms of use apply. Rights reserved

9
Vol.:(0123456789)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
While coal-red power plants have historically wreaked havoc on vegetation ecosystems, exacerbating habitat
loss and degradation, the loss of economic value of carbon sinks reects the negative spillovers caused by wind
power, which calls for more attention to the impact of green energy on regional ecology while expanding the wind
power industry and developing clean energy. To mitigate the potential threats of renewable energy deployment to
vegetation ecosystems, a comprehensive monitoring system is needed to assess the impacts of wind farms on plant
communities over time. Regular vegetation surveys, remote sensing, and other monitoring techniques should be
used to track changes in vegetation succession and distribution. Landscape planning should be conducted before
wind farm construction to identify areas of particular importance for plant biodiversity and to avoid placing
wind farms in those areas. Oset measures, such as reforestation, habitat restoration, and the creation of wildlife
corridors, can be implemented to mitigate the negative externalities of wind farm development. However, it’s
crucial to contextualize these ndings within the broader initiative to transition from fossil fuels to renewable
energy, a move that comes with its own set of signicant environmental benets. Future research should examine
the ecological requirements of plant species and develop balanced mitigation strategies for energy transition.
Methods
Data
e data primarily consists of two parts. e rst part includes wind power data, collected from the National
and Provincial Development and Reform Commission, the Energy Bureau, wind power project public bidding
documents, wind power project completion environmental impact acceptance reports, and the China Clean
Development Mechanism’s ocial website. e wind power data includes all completed and operational wind
farms as of 2022 at the county level, with detailed information on construction completion dates, installed
capacity, and the longitude and latitude coordinates of the center of each wind farm. We collected data from
2404 wind farms built between 1994 and 2022, as shown in Fig.1. Wind farms with incomplete information on
construction time and installation data were excluded from the nal sample.
We utilized the geographic information system (GIS) to determine the land surface area covered by the
turbines of each wind farm in a two-step process. Firstly, we geocoded the latitude and longitude of each wind
farm by combining its address with high-resolution remote sensing images. We determined the exact location of
each wind farm based on the recorded address and the visibility of wind turbines in the remote sensing images.
Secondly, we divided the national land surface into 2km × 2km grids. Because the typical distance between two
turbines is 0.5km, and wind turbines are usually equally distributed within a given wind farm, we calculated
the number of grids occupied by the wind turbines by dividing the total number of turbines in each wind farm
by 16. e grids closest to the wind farm’s location are designated as the land parcels occupied by the wind farm,
and we considered the outward boundaries of the grids as the spatial boundaries of each wind farm.
e second dataset includes plant biomass production (PBP) indicators from Google Earth Engine. Ten indi-
ces were used to measure PBP, including NDVI, EVI, FPAR, LAI, GPP, NP, NPP, PTC, PNTV, and PNV. NDVI
and EVI assess vegetation growth and cover, with values ranging from − 1 to 1. FPAR measures vegetation’s light
energy eciency, with values from 0 to 1 (normalized to [0, 100] in analysis). LAI indicates vegetation growth and
canopy structure, with values from 0 to 1. GPP and NP represent organic carbon xation and net photosynthesis,
respectively. NPP measures organic carbon remaining in plants, with GPP and NPP values typically ranging
from 0 to 0.3. PTC, PNTV, and PNV describe the percentage of tree cover, non-tree vegetation cover, and non-
vegetation cover, respectively, with values from 0 to 100. Supplementary Table1 provides detailed denitions.
e Terra satellite provides data for the years 2000–2022, using the MODIS (Moderate Resolution Imag-
ing Spectroradiometer) remote sensing data product. We measure NDVI and EVI using the 16-day synthetic
vegetation index product MOD13Q1.061, while FPAR and LAI use the 8-day synthetic land leaf area index
product MOD15A2H.061. Similarly, we measure GPP and NP using the 8-day synthetic primary productiv-
ity product MOD17A2H.006, while NPP is measured using the 1-year synthetic primary productivity prod-
uct MOD17A3HGF.006. To determine the coverage rate of planting areas, we use the annual data product
MOD44B.006 to measure the PTC, PNTV, and PNV indices. Since the initial data covers a global scale, we use
the Java Script programming platform to spatially locate the longitude and latitude coordinates of the centers
of the 2,404 wind farms and obtain the monthly vegetation data based on specic wind farm locations within
each buer zone.
Empirical strategy
We divide the study area within a 20km from the center of each wind farm into 12 buer zones based on distance,
including 1km, 1–2km, 2–3km, 3–4km, 4–5km, 5–6km, 6–7km, 7–8km, 8–9km, 9–10km, 10–15km, and
15–20km from the center of the wind farm, resulting in a total of 3288month-by-buer-zone observations (i.e.
274months from January 2000 to October 2022 multiplied by 12 buer zones) for a single wind farm. To elimi-
nate the eects of multiple turbine expansions on the analysis, we limit our consideration to 1,936 wind farms
that were built aer 2002 and have only undergone a single installation (Supplementary Table2). Consequently,
we have a total of 6,365,568 observations for analysis.
Fixed eects approach
We implement a high-dimensional xed eects approach to estimate the average eect of wind farm construction
on PBP under a certain distance buer zone, specied as follows:
(2)
Yk,d
it
=θk,d
postit
+X′
it
γk,d+δ
j
×η
y
+τ
m
+µ
g
+ε
it ,
Content courtesy of Springer Nature, terms of use apply. Rights reserved

10
Vol:.(1234567890)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
where
Yk,d
it
denotes one of the
k
PBP indicators of a distance buer zone
d
of wind farm
i
at month
t
.
postit
equals
1 if it is aer the wind farm installation and 0 otherwise.
X′
it
is a vector of control variables that may aect PBP,
including the number of wind turbines, wind power capacities of the wind farm, precipitation, temperature at
two meters above the ground, nighttime light, and elevation. In particular, precipitation and temperature are
good measures of surrounding natural or climatic conditions to exclude the eects of global warming issues such
as extreme weather events on the vegetation growth; nighttime light data has garnered considerable attention
in recent scientic literature as a robust and informative proxy for assessing local economic development70–72.
is approach allows researchers to capture the extent of anthropogenic activity and eectively control for its
associated implications on human–environment interactions73,74, particularly concerning vegetation around wind
farms. Supplementary Table1 provides detailed descriptions of all control variables.
θk
,
d
is the main coecient of
interest, which measures the dierence of mean values of a PBP indicator
k
before and aer wind farm installa-
tion within buer zone
d
, and
γk
,
d
denotes the vector of coecients for the control variable
X
. In the presented
equation, county-by-year xed eects (
δj×ηy
) not only represent unobserved attribute changes within the
counties (j) housing wind farms but also precisely capture and control the economic dynamics of these regions.
ese dynamics encompass, but are not limited to, variations in Gross Domestic Product, industrial output, and
trac indicators. e comprehensiveness and granularity of these xed eects eectively mitigate any natural
or anthropogenic confounding factors in the estimation of wind farm impacts. Consequently, their incorpora-
tion into the study not only bolsters condence in the accuracy of the estimated benets of wind farms but also
ensures the reliability and precision of the analysis in identifying the impacts of wind farms on local economies
and environments.
τm
represents month-of-year xed eects, capturing any monthly characteristics or seasonal
eects such as changes in local wind speed and direction.
µg
represents land type xed eects, which control for
unobservables under dierent topographic features. We considered seven land type categories, including the
plain, terrace, hill, low relief mountain, intermediate relief mountain, high relief mountain, and extremely high
relief mountain.
εit
is the idiosyncratic error term. We repeatedly estimated Eq.(2) in the methods section for
all distance buer zones (
d
=
1, 2, ..., 12
) and for all PBP indicators (
k
=
1, 2, ..., 10
). All regressions employ
identical specications. We also clustered the standard errors at the wind farm level, which accounts for correla-
tions between observations around the same wind farm.
Dynamic treatment eect
To examine the dynamic eects of wind farm construction, based on Eq.(2), we replace the dummy
postit
with
a set of dummy variables indicating dierent time periods to explicitly estimate the impact of wind farms six
months before, the rst year, the second year, the third year, and more than three years aer the installation of
the wind farm. e dynamic model is specied as follows:
where
Yk,d
it
denotes one of the
k
PBP indicators of a distance buer zone
d
of wind farm
i
at month
t
.
lag1
is a
dummy that equals 1 if it is at least six months before installation and 0 otherwise;
leadq
is a dummy indicating
one year (
q
=
1
), two years (
q
=
2
), three years (
q
=
3
), and more than three years (
q
=
4
) aer installation.
θk,d
−1
and
θk,d
+
q
’s are coecients measuring the dynamic impacts of wind farm installation. e remaining terms are
identical to Eq.(2). Since NPP, PTC, PNTV, and PNY are measured by year,
lag1
indicates one year, other than
half a year, before installation for the regressions of these four indicators. For all regressions, “more than half a
year (or one year) before installation” serves as the reference group.
Heterogeneous eects
We also estimate Eq.(4) below with dierent groups of subsamples to explore the heterogeneous eects across
season, elevation, and land type.
where
Yk,d,h
it
denotes one of the
k
PBP indicators within a distance buer zone
d
of wind farm
i
at month
t
.
h
denotes one of the three groups of subsamples. For seasonal heterogeneous eects (
h=1
), we considered four
subsamples—spring (March to May), summer (June to August), autumn (September to November), and winter
(December to February), which results in 480 (10 × 12 × 4) regressions and
θk,d,h
’s. For heterogeneous eects
of elevation (
h=2
), we considered three subsamples—elevation less than 500m, elevation between 500 and
1,500m, and elevation larger than 1,500m, which results in 360 (10 × 12 × 3) regressions and
θk,d,h
’s. For hetero-
geneous eects of land type (
h=3
), we considered ve subsamples—plain, terrace, hill, low relief mountain,
intermediate relief mountain, resulting in 600 (10 × 12 × 5) regressions and
θk,d,h
’s.
We determine which subsample of the elevation group and land type group each wind farm belongs to by the
mean values of elevation (DEM) and topographic (Geomor) indicators within the 1km buer zone for the wind
farm. For the seasonal subsample, it is determined directly from the specic month recorded for each observa-
tion. In addition, in the land type group, the ve available subsamples did not cover the entire sample. Due to
the small number of wind farms distributed in the two types of land type—high relief mountain and extremely
high relief mountain—we did not include estimates of these two subsamples.
(3)
Y
k,d
it =θk,d
−1lag1it +
4

q
=1
θk,d
+qleadqit +X′
it γk,d+δj×ηy+τm+µg+εit
,
(4)
Yk,d,h
it
=θ
k,d,h
post
it
+X′
it
γ
k,d,h
+δ
j
×η
y
+τ
m
+µ
g
+ε
it ,
Content courtesy of Springer Nature, terms of use apply. Rights reserved

11
Vol.:(0123456789)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
Data availability
e datasets used and/or analysed during the current study are available from the corresponding author on
reasonable request.
Received: 29 September 2023; Accepted: 11 December 2023
References
1. Garcia, R. A., Cabeza, M., R ahbek, C. & Araújo, M. B. Multiple dimensions of climate change and their implications for biodiversity.
Science 344(6183), 1247579 (2014).
2. Fankhauser, S. et al. e meaning of net zero and how to get it right. Nature Clim. Change 12(1), 15–21 (2022).
3. Schreifels, J. J., Fu, Y. & Wilson, E. J. Sulfur dioxide control in China: policy evolution during the 10th and 11th Five-year Plans
and lessons for the future. Energy Policy 48, 779–789 (2012).
4. Driscoll, C. T., Driscoll, K. M., Mitchell, M. J. & Raynal, D. J. Eects of acidic deposition on forest and aquatic ecosystems in New
York State. Environ. Pollut. 123(3), 327–336 (2003).
5. Wang, S. X. et al. Mercury emission and speciation of coal-red power plants in China. Atmos. Chem. Phys. 10(3), 1183–1192
(2010).
6. Smith, P. et al. e role of ecosystems and their management in regulating climate, and soil, water and air quality. J. Appl. Ecol.
50(4), 812–829 (2013).
7. Rai, P. K. Impacts of particulate matter pollution on plants: Implications for environmental biomonitoring. Ecotoxicol. Environ.
Saf. 129, 120–136 (2016).
8. Dale, S. BP Statistical Review of World Energy.BP Plc, London, United Kingdom 14–16 (2021).
9. Tracking SDG 7: e Energy Progress Report(IEA, IRENA, UNSD, World Bank & WHO, 2020).
10. Girardin, C. A. et al. Nature-based solutions can help cool the planet—if we act now. Nature 593(7858), 191–194 (2021).
11. Mori, A. S. et al. Biodiversity–productivity relationships are key to nature-based climate solutions. Nature Clim. Change 11(6),
543–550 (2021).
12. Rahman, M. M. et al. Co-benets of protecting mangroves for biodiversity conservation and carbon storage. Nature Commu.
ications 12(1), 3875 (2021).
13. aker, M., Zambre, A. & Bhosale, H. Wind farms have cascading impacts on ecosystems across trophic levels. Nature Ecol. Evol.
tion 2(12), 1854–1858 (2018).
14. Gibson, L., Wilman, E. N. & Laurance, W. F. How green is ‘green’energy?. Trends Ecol. Evol. 32(12), 922–935 (2017).
15. Causon, P. D. & Gill, A. B. Linking ecosystem services with epibenthic biodiversity change following installation of oshore wind
farms. Environ. Sci. Policy 89, 34 (2018).
16. De Mesel, I., Kerckhof, F., Norro, A., Rumes, B. & Degraer, S. Succession and seasonal dynamics of the epifauna community on
oshore wind farm foundations and their role as stepping stones for non-indigenous species. Hydrobiologia 756(1), 37–50 (2015).
17. Sayed, E. T. et al. A critical review on environmental impacts of renewable energy systems and mitigation strategies: Wind, hydro,
biomass and geothermal. Sci. Total Environ. 766, 144505 (2021).
18. Kumara, H. N. et al. Responses of birds and mammals to long-established wind farms in India. Sci. Rep. 12(1), 1–15 (2022).
19. Xu, K. et al. Positive ecological eects of wind farms on vegetation in China’s Gobi desert. Sci. Rep. 9(1), 6341 (2019).
20. Gușatu, L. F. et al. Spatial and temporal analysis of cumulative environmental eects of oshore wind farms in the North Sea basin.
Sci. Rep. 11(1), 1–18 (2021).
21. Sonter, L. J., Dade, M. C., Watson, J. E. M. & Valenta, R. K. Renewable energy production will exacerbate mining threats to biodi-
versity. Nature Commun. 11(1), 1–6 (2020).
22. Spillias, S., Kareiva, P., Ruckelshaus, M. & McDonald-Madden, E. Renewable energy targets may undermine their sustainability.
Nature Clim. Change 10(11), 974–976 (2020).
23. Dirnböck, T. et al. Climate and air pollution impacts on habitat suitability of Austrian forest ecosystems. PloS one 12(9), e0184194
(2017).
24. Vautard, R. et al. Regional climate model simulations indicate limited climatic impacts by operational and planned European wind
farms. Nature Commun. 5(1), 1–9 (2014).
25. Zhou, L. et al. Impacts of wind farms on land surface temperature. Nature Clim. Change 2(7), 539–543 (2012).
26. Li, Y. et al. Climate model shows large-scale wind and solar farms in the Sahara increase rain and vegetation. Science 361(6406),
1019–1022 (2018).
27. Kane, D. T. Microclimate eects of wind farms on local crop yields. J. Environ. Econ. Manage. 96, 159–173 (2019).
28. Roscher, C. et al. Using plant functional traits to explain diversity–productivity relationships. PloS one 7(5), e36760 (2012).
29. Abbasi, S. A. & Abbasi, N. e likely adverse environmental impacts of renewable energy sources. Appl. Energy 65(1–4), 121–144
(2000).
30. Liu, Z., Li, G. & Wang, G. Can wind farms change the phenology of grassland in China?. Sci. Total Environ. 832, 155077 (2022).
31. Nazir, M. S., Ali, N., Bilal, M. & Iqbal, H. M. N. Potential environmental impacts of wind energy development: A global perspec-
tive. Curr. Opin. Environ. Sci. Health 13, 85–90 (2020).
32. Zhou, L., Tian, Y., Baidya Roy, S., Dai, Y. & Chen, H. Diurnal and seasonal variations of wind farm impacts on land surface tem-
perature over western Texas. Clim. Dyn. 41(2), 307–326 (2013).
33. Hoang, N. T. & Kanemoto, K. Mapping the deforestation footprint of nations reveals growing threat to tropical forests. Nature
Ecol. Evol. 5(6), 845–853 (2021).
34. Zhang, H., Li, Y. & Zhu, J. K. Developing naturally stress-resistant crops for a sustainable agriculture. Nature Plants 4(12), 989–996
(2018).
35. McElroy, M. B., Lu, X., Nielsen, C. P. & Wang, Y. Potential for wind-generated electricity in China. Science 325(5946), 1378–1380
(2009).
36. Sherman, P., Chen, X. & McElroy, M. Oshore wind: An opportunity for cost-competitive decarbonization of China’s energy
economy. Sci. Adv. 6(8), eaax9571 (2020).
37. Xu, M., David, S."China doubles new renewable capacity in 2020; still builds thermal plants".Reuters. Retrieved11 March2021
(2021).
38. Global Oshore Wind Report 2022 (GWECs, 2022); https:// gwec. net/ wp- conte nt/ uploa ds/ 2022/ 06/ GWEC- Global- Osh ore-
Wind- Report- 2022. pdf
39. China Built More Oshore Wind In 2021 an Every Other Country Built In 5 Years (Forbes, 2022); https:// www. forbes. com/
sites/ david rvett er/ 2022/ 01/ 26/ china- built- more- osh ore- wind- in- 2021- than- every- other- count ry- built- in-5- years/? sh= 15e32
d3546 34
40. Hepburn, C. et al. Towards carbon neutrality and China’s 14th Five-Year Plan: Clean energy transition, sustainable urban develop-
ment, and investment priorities. Environ. Sci. Ecotechnol. 8, 100130 (2021).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

12
Vol:.(1234567890)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
41. Li, B. et al. How to strive for balance of coastal wind energy development with waterbird conservation in the important coastal
wetlands, a case study in the Chongming Islands of East China. J. Clean. Prod. 263, 121547 (2020).
42. Tang, B. et al. e observed impacts of wind farms on local vegetation growth in northern China. Remote Sens. 9(4), 332 (2017).
43. Wu, X.-L. et al. Satellite-based assessment of local environment change by wind farms in China. Earth Space Sci. 6(6), 947–958
(2019).
44. Carr-Harris, A. & Lang, C. Sustainability and tourism: e eect of the United States’ rst oshore wind farm on the vacation
rental market. Resource Energy Econ. 57, 51–67 (2019).
45. Wang, G., Li, G. & Liu, Z. Wind farms dry surface soil in temporal and spatial variation. Sci. Total Environ. 857, 159293 (2023).
46. Wang, C., Fu, B., Zhang, L. & Xu, Z. Soil moisture–plant interactions: an ecohydrological review. J. Soils Sedim. 19, 1–9 (2019).
47. Berg, A., Sheeld, J. & Milly, P. C. Divergent surface and total soil moisture projections under global warming. Geophys. Res. Lett.
44(1), 236–244 (2017).
48. Schnorr, S. L. et al. Gut microbiome of the Hadza hunter-gatherers. Nature Commun. 5(1), 3654 (2014).
49. Pryor, S. C., Barthelmie, R. J. & Shepherd, T. J. 20% of US electricity from wind will have limited impacts on system eciency and
regional climate. Sci. Rep. 10(1), 541 (2020).
50. Song, D. et al. Annual energy production estimation for variable-speed wind turbine at high-altitude site. J. Modern Power Syst.
Clean Energy 9(3), 684–687 (2020).
51. Pătru-Stupariu, I. et al. Do wind turbines impact plant community properties in mountain region?. Biologia 74, 1613–1619 (2019).
52. Barr y, G., Peter, B. & Bruno, M. Review: Wind impacts on plant growth, mechanics and damage. Plant Sci. 245, 94–118 (2016).
53. Moghaddam, S. Z. Generation and transmission expansion planning with high penetration of wind farms considering spatial
distribution of wind speed. Int. J. Electr. Power Energy Syst. 106, 232–241 (2019).
54. Bolívar-Cimé, B., Bolívar-Cimé, A., Cabrera-Cruz, S. A., Muñoz-Jiménez, Ó. & Villegas-Patraca, R. Bats in a tropical wind farm:
species composition and importance of the spatial attributes of vegetation cover on bat fatalities. J. Mammal. 97(4), 1197–1208
(2016).
55. Gao, X. et al. Comparative experimental investigation into wake characteristics of turbines in three wind farms areas with varying
terrain complexity from LiDAR measurements. Appl. Energy 307, 118182 (2022).
56. Kozmar, H., Allori, D., Bartoli, G. & Borri, C. Wind characteristics in wind farms situated on a hilly terrain. J. Wind Eng. Ind.
Aerodyn. 174, 404–410 (2018).
57. Abdeslame, D., Merzouk, N. K., Mekhtoub, S., Abbas, M. & Dehmas, M. Estimation of power generation capacities of a wind farms
installed in windy sites in Algerian high plateaus. Renew. energy 103, 630–640 (2017).
58. Margreth, A. Origins of low-relief plateaus. Nature Geosci. 10(8), 541–542 (2017).
59. Lira-Martins, D. et al. Soil properties and geomorphic processes inuence vegetation composition, structure, and function in the
Cerrado Domain. Plant Soil 476(1–2), 549–588 (2022).
60. Li, J. et al. Landform-related permafrost characteristics in the source area of the Yellow River, eastern Qinghai-Tibet Plateau.
Geomorphology 269, 104–111 (2016).
61. Nicoll, T., Brierley, G. & Yu, G. A. A broad overview of landscape diversity of the Yellow River source zone. J. Geograph. Sci. 23,
793–816 (2013).
62. Pan, Y. et al. A large and persistent carbon sink in the world’s forests. Science 333, 988–993. https:// doi. org/ 10. 1126/ scien ce. 12016
09 (2011).
63. Favero, A., Mendelsohn, R. & Sohngen, B. Using forests for climate mitigation: sequester carbon or produce woody biomass?.
Clim. Change 144, 195–206. https:// doi. org/ 10. 1007/ s10584- 017- 2034-9 (2017).
64. Nunes, L. J. R., Meireles, C. I. R., Pinto Gomes, C. J. & Almeida Ribeiro, N. Forest contribution to climate change mitigation:
Management oriented to carbon capture and storage. Climate https:// doi. org/ 10. 3390/ cli80 20021 (2020).
65. omas, S. C. & Martin, A. R. Carbon content of tree tissues: a synthesis. Forests 3, 332–352. https:// doi. org/ 10. 3390/ f3020 332
(2012).
66. Fang, J. et al. Forest biomass carbon sinks in East Asia, with special reference to the relative contributions of forest expansion and
forest growth. Glob. Change Biol. 20(6), 2019–2030. https:// doi. org/ 10. 1111/ gcb. 12512 (2014).
67. Holst, J. et al. Impacts of summer water limitation on the carbon balance of a Scots pine forest in the southern upper Rhine plain.
Agric. For. Meteorol. 148(11), 1815–1826 (2008).
68. Ceccherini, G. et al. Abrupt increase in harvested forest area over Europe aer 2015. Nature 583, 72–77. https:// doi. org/ 10. 1038/
s41586- 020- 2438-y (2020).
69. Zeng, W., Tang, S. Modeling compatible single-tree aboveground biomass equations of masson pine (Pinus massoniana) in South
China. Nat Prec (2012). https:// www. nature. com/ artic les/ npre. 2012. 6754.1
70. Gibson, J., Olivia, S., Boe-Gibson, G. & Li, C. Which night lights data should we use in economics, and where?. J. Dev. Econ. 149,
102602 (2021).
71. Levin, N. et al. Remote sensing of night lights: A review and an outlook for the future. Remote Sens. Environ. 237, 111443 (2020).
72. Gibson, J., Olivia, S. & Boe-Gibson, G. Night lights in economics: Sources and uses 1. J. Econ. Surv. 34(5), 955–980 (2020).
73. Goldblatt, R., Heilmann, K. & Vaizman, Y. Can medium-resolution satellite imagery measure economic activity at small geogra-
phies? Evidence from Landsat in Vietnam. World Bank Econ. Rev. 34(3), 635–653 (2020).
74. Chen, X. & Nordhaus, W. D. Using luminosity data as a proxy for economic statistics. Proc. Natl. Acad. Sci. 108(21), 8589–8594
(2011).
Acknowledgements
is research was supported by the National Natural Science Foundation of China (No. 72373162 and No.
72173109).
Author contributions
L.G. and Y.M. conceived the original concept and ideas; J.Q., Y.Y. and L.M. collected and analysed the data;
Q.W., J.Q. and Y.Y. nished the literature review; L.G. and Y.Y. performed the formal analysis; L.G., Q.W. and
Y.M. realised the maps and graphs; L.G., Q.W. wrote the rst dra; P.L. provided supervision. All authors read
and approved for submission.
Competing interests
e authors declare no competing interests.
Additional information
Supplementary Information e online version contains supplementary material available at https:// doi. org/
10. 1038/ s41598- 023- 49650-9.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

13
Vol.:(0123456789)
Scientic Reports | (2023) 13:22366 | https://doi.org/10.1038/s41598-023-49650-9
www.nature.com/scientificreports/
Correspondence and requests for materials should be addressed to Y.M.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access is article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
© e Author(s) 2023
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com