Next Article in Journal
Wood Properties Characterisation of Thermo-Hydro Mechanical Treated Plantation and Native Tasmanian Timber Species
Next Article in Special Issue
Estimating VAIA Windstorm Damaged Forest Area in Italy Using Time Series Sentinel-2 Imagery and Continuous Change Detection Algorithms
Previous Article in Journal
The Development of Pneumatic Fatigue Test Rig for Wood-Based Specimens
Previous Article in Special Issue
Remote Sensing Applied in Forest Management to Optimize Ecosystem Services: Advances in Research
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Forests
Volume 11
Issue 11
10.3390/f11111188
forests-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Fifty Years of Change in a Coniferous Forest in the Qilian Mountains, China—Advantages of High-Definition Remote Sensing

by
Shu Fang
1 and
Zhibin He
2,*
1
College of Urban, Rural Planning and Architectural Engineering, Shangluo University, Shangluo 726000, China
2
Linze Inland River Basin Research Station, Chinese Ecosystem Research Network, Key Laboratory of Eco-hydrology of Inland River Basin, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou 730000, China
*
Author to whom correspondence should be addressed.
Forests 2020, 11(11), 1188; https://doi.org/10.3390/f11111188
Submission received: 9 October 2020 / Revised: 5 November 2020 / Accepted: 9 November 2020 / Published: 10 November 2020
(This article belongs to the Special Issue Analysis of Forest Landscape and Land-Use Based on Remote Sensing Technology)
Browse Figures
Review Reports Versions Notes

Abstract

:
Mountain ecosystems are significantly affected by climate change. However, due to slow vegetation growth in mountain ecosystems, climate-induced vegetation shifts are difficult to detect with low-definition remote sensing images. We used high-definition remote sensing data to identify responses to climate change in a typical Picea crassifolia Kom. forest in the Qilian Mountains, China, from 1968 to 2017. We found that: (1) Picea crassifolia Kom. forests were distributed in small patches or strips on shaded and partly shaded slopes at altitudes of 2700–3250 m, (2) the number, area, and concentration of forest patches have been increasing from 1968 to 2017 in relatively flat and partly sunny areas, but the rate of area increase and ascend of the tree line slowed after 2008, and (3) the establishment of plantation forests may be one of the reasons for the changes. The scale of detected change in Picea crassifolia Kom.forest was about or slightly below 30 m, indicating that monitoring with high-resolution remote sensing data will improve detectability and accuracy.

1. Introduction

The temperature increase associated with global climate change was three times higher in mountain areas in the past 40 years than the global average, and the frequency of extreme climate events was also significantly greater in mountainous than in other ecosystems [1]. The tree line is the continuous forest boundary of a mountain forest, and it is the transition zone from forest to tundra [2,3,4]. Mountain vegetation, especially at the alpine tree line, is highly sensitive to climate change and can reflect climate change more quickly than other types of vegetation [5]. In fact, the upper limit of the tree line is expected to shift up by as much as 300 to 600 m in elevation as a result of climate warming [6]. Further, deciduous forests may increase in presence at the alpine tree line due to warming, increase in length of the growing season, and decreased severity of the winter environment [7].
However, the upward movement of the alpine tree line caused by a temperature rise is relatively slow [8]; additionally, warming has been linked to drought stress at the alpine tree line, which may decrease the rate of forest growth [9]. The impact of the warming climate on coniferous forests was not significant because snow, wind damage, and frequency of pests and diseases were delayed in alpine areas of the Northern Hemisphere until the 1980s [10].
The Qilian Mountains are located in the arid northwestern region of China and play an important role in contributing water to the lowlands. Forest vegetation in the Qilian Mountains is a valuable forest resource with the ecological function of water conservation [11].
Picea crassifolia Kom. (Qinghai spruce) is the dominant tree species in the Qilian Mountains, forming an almost pure spruce forest in places [12]. Picea crassifolia Kom. has been widely used in dendroclimatology to determine the relationships between tree growth in the Qilian Mountains and climatic factors [13,14,15]. A recent survey showed that the Picea crassifolia Kom. tree line trended upward [16], but the sample size in the survey was limited. In efforts to determine the responses of Qinghai spruce forest to climate change, traditional dendroclimatology approaches result in great uncertainty and inconsistency with sample surveys. At the same time, research scale is not representative, remaining on a single tree. Thus, a larger-scale approach is needed to determine the response of Picea crassifolia Kom. forests to climate warming.
In mountainous areas where data are scarce and sites may be inaccessible, remote sensing data and GIS (Geographic Information System) technology are particularly useful in monitoring changes in forest location and landscape patterns [17,18]. Continuity in remote sensing data allows for dynamic monitoring of the tree line position and pattern change [19,20]. As early as 1995, tree lines of patch forest ecotones in the Rocky Mountain National Park were detected with aerial photographs of 1988 and 1990 [21]. The tree line can be directly detected based on the vegetation type over the remote sensing image, and the identification of the tree line can be more accurate with the use of the Normalized Difference Vegetation Index (NDVI) or topographic data such as elevation and slope [22,23]. Direct classification for Landsat images can be used for extracting tree line ecotone, but a complex method, such as linear spectral mixture analysis, should be used for detecting the alpine tree lines [24,25].
Changes in the tree line using remote sensing data are mainly identified with vegetation coverage or NDVI [26]. However, coniferous forests changed more slowly than tundra and shrubland in alpine tree line areas [26], and slow vegetation changes may be difficult to detect with low-resolution remote sensing data. Tree lines in coniferous forests in the Finnish Lapland have increased in tree density, and expanded upward in the last three decades, but that could not be detected with NDVI or land-cover classification from Landsat [27]. High-resolution remote sensing data make it possible to monitor some areas where the tree line does not change significantly even at species level, and where accuracy and precision of monitoring results will be improved [28]. The growth rate of Picea crassifolia Kom. natural forests in the Qilian Mountains is slow, and the natural renewal capacity of secondary forests is low [29], making it difficult to monitor forest changes using general remote sensing images, such as Landsat with a resolution of 30 m.
To improve monitoring of changes in Picea crassifolia Kom. forest, we chose high-resolution aerial photography and remote sensing images of the Dayekou watershed in the Qilian Mountains at a resolution of about 2 m for 1968, 2008, and 2017. We aimed to determine whether (1) the Picea crassifolia Kom. tree line in the Qilian Mountains changed in the past 50 years, (2) the landscape pattern of the tree line position has changed, and (3) changes were consistent with climate change. The results of this study provide a reference for high-resolution remote sensing image monitoring of the position and landscape pattern of forest tree lines for slowly growing coniferous forests in alpine mountain areas.

2. Materials and Methods

2.1. Study Area

We conducted this study at the Dayekou watershed (38°26′–38°35′ N, 100°14′–100°19′ E) in the upper reaches of the Heihe River Basin in the Qilian Mountains of China (Figure 1). It is located in the Xishui Forest District of the middle Qilian Mountains, and has an area of about 73 km2 at an altitude of 2500–4700 m. Slopes of the low-mountain areas in the Dayekou watershed are between 10 and 30°, and those of the high-mountain areas are about 40°. Climate in the Dayekou watershed is dry and semi-humid, with average annual temperature of 2.0–3.5 °C, and average July temperature of 10–14 °C. Annual precipitation is 350–450 mm, which is concentrated in June–September, annual evaporation is 1050–1100 mm, annual cumulative duration of sunshine is 1890–1900 h, and average annual relative humidity is 60% to 65%. The foundation species is Picea crassifolia Kom., which has patchy distribution on shaded and semi-shaded slopes at an altitude of 2500–3400 m. The areas also contain sparse Qilian juniper (Sabina przewalskii Kom.). The soil is mainly grey cinnamon and chestnut, and is characterized as relatively thin, with mainly silt sand texture [30].

2.2. Remote-Sensing Data Acquisition and Pre-Processing

We chose high-resolution remote sensing images of the Dayekou watershed for 1968, 2008, and 2017. Specifically, we used the 1968 keyhole historical aerial photograph data [31] with a resolution of 1.8 m, the 2008 Quickbird image data [32] with a resolution of 2.4 m, and the 2017 ZY-3 satellite images [33] with a resolution of 2.1 m. The images had the least cloud amounts and high data quality and were selected among satellite images of similar resolution for the same period.
Next, we selected data for accuracy verification and digital elevation model (DEM) data from the Heihe Data Management Center (http://www.heihedata.org/) (Figure 2). First, we used the following datasets to establish a database of field sampling points to verify the accuracy of remote sensing image interpretation: (1) a dataset of forest structure at the fixed sampling plot in the Pailugou watershed and Dayekou watershed foci experiment area from 2003 [34], (2) a dataset of forest structure at the fixed sampling plot in the Pailugou watershed and Dayekou watershed foci experiment area for 2007 [35], (3) a dataset of forest structure at the temporary forest sampling plot in the Dayekou watershed foci experimental area for 2008 [36], (4) the Heihe Integrated Remote Sensing Joint Test, Arid Region Hydrological Experimental Area and Forest Hydrological Experimental Area, Land Use and Land Cover Survey Dataset [37], and (5) a dataset of forest structure measurements for the fixed forest sampling plots in the Dayekou and Pailugou watershed foci experimental areas (2003–2007) [38]. Then, we used the 1 m digital elevation model (DEM) of the Dayekou watershed established by wordview-2 data in 2012 [39] as the basis for extracting relevant terrain factors.
To improve accuracy of remote sensing classification, we used drone image photographed by DJI (Da Jiang Innovations) phantom 4 pro for areas with plantations and Qilian Juniper forests (Figure 3).
After pre-processing of the three images for atmospheric interference, and georeference calibration, we used the k-means unsupervised classification method in ENVI 5.3 [40] to classify patchy vegetation features more accurately [41], and to divide the Dayekou watershed into coniferous and non-coniferous forests. Some Qilian juniper trees in the coniferous forest were eliminated by visual adjustment based on forest canopy closure, distribution, location, and data from drone surveys; then, the remaining coniferous forests were classified as Picea crassifolia Kom. The accuracy of the decoded image was verified based on field and drone sampling points of the Heihe Data Center. The kappa index was 0.8630, and total accuracy was 93.67%. Finally, the distribution maps of Picea crassifolia Kom. forests in the Dayekou watershed for the 1968, 2008, and 2017 were obtained, and the boundary of the forest is the tree line. We have counted the maximum, minimum, and average values of each forest patch to explore the upper, lower, and overall tree line changes.

2.3. Data Analysis

2.3.1. Changes in the Distribution of Picea crassifolia Kom.

Regional statistics and spatial analysis for the three phases were performed with ESRI ArcMap 10.4 [42] raster data analysis to obtain characteristics of Picea crassifolia Kom. distribution. Topographic data were superimposed on the distribution data, and elevation, slope, and aspect characteristics of Picea crassifolia Kom. distribution were obtained for the three-phase remote sensing images.
The area conversion matrix for 1968, 2008, and 2017 remote sensing images of Picea crassifolia Kom. forest and non-Picea crassifolia Kom. forest was first calculated to describe the area change between Qinghai spruce forest and non-Picea crassifolia Kom. forest between time periods.
We used three terrain indicators, elevation, slope, and aspect, which were calculated by the ‘Surface’ tool in ‘Spatial Analyst Tools’ in ESRI ArcMap 10.4, to detect dynamic changes in forest distribution. Then we used ‘Reclassify’ in ‘Spatial Analyst Tools’ to classify elevation, slope, and aspect. The forest was distributed at elevations between 2600 and 3550 m, so we divided elevation into 19 levels, with the first level at below 2600 m, one level added every 50 m, and the last level at >3450 m. Slope ranged from 0 to 83.99°, and we divided slope into eighteen 5° increments (categories). The aspect distribution of forests was mainly related to shaded and sunny slopes, so the slope direction was divided into 5 categories, with 0 for no slope direction, 1 for shaded slopes (N (0–22.5°, 337.5–360°)), 2 for semi-shaded (NE (22.5–67.5°), E (67.5–112.5°), NW (292.5–337.5°)), 3 for sunny (S (157.5–202.5°)), and 4 for partly-sunny slopes (SW (112.5–157.5°), SE (202.5–247.5°), and W (247.5–292.5°)).
Subsequently, we used the terrain distribution index P of elevation, slope, and aspect (Equation (1)) to analyze the relationship between distribution and topography of Picea crassifolia Kom. forest, and vegetation change rate index K (Equation (2)) to determine the rate of change of Picea crassifolia Kom. forest area between different periods.
P = S ie / S e ,
K = S b − S a / S a ) × 1 / t ,
where: e represents topographic factors of elevation, slope, and aspect, S ie represents the area of Picea crassifolia Kom. forest at a specific level in e elevation, slope, and aspect, S e is the total area in a specific level of e elevation, slope, and aspect within the entire study area, S a and S b   are the area of vegetation at the beginning and end of the study period, and t is the length of the study period.

2.3.2. Landscape Changes of Picea crassifolia Kom.

Landscape index was used to describe landscape patterns and their changes, and to establish the relationship between landscape pattern and process, as it is the most commonly used quantitative research method in landscape ecology [43]. To calculate the landscape pattern index, the scale of calculation is determined first. Based on the results of extensive landscape-scale research [44,45], and conditions of the selected research area, 20 scales were selected from 1 to 30 m, with a 1 m interval between 1 and 10 m, and a 2 m interval between 10 and 20 m. We calculated the landscape pattern index at different scales in FRAGSTSTS 4.2 [46] to obtain the response map at different landscape pattern scales. We selected an optimal scale for our study, and, at this scale, we calculated the landscape pattern index of Picea crassifolia Kom. and analyzed changes within it for the last 50 years. In this study, we calculated six landscape pattern indices (Table 1) for Picea crassifolia Kom. forest at the landscape level [47,48,49,50].

3. Results

3.1. Shift in the Forest Cover

The Picea crassifolia Kom. forest in the Dayekou watershed was distributed in patches and strips (Figure 4). Most patches were distributed in the northeast-southwest direction and were found mostly in the north of the basin. Overall, the number and area of Picea crassifolia Kom. patches in the basin have been increasing since 1968: there were 222 forest patches in 1968, 263 in 2008, and 264 in 2017. Forest area in 1968 was 12.44 km2, accounting for 17.04% of the entire basin area. By 2008, forest area increased to 14.94 km2, and the proportion of total forest area has increased by 3.43%. By 2017, forest area reached 15.48 km2, accounting for 21.21% of the total watershed area. The average area of forest patches in Dayekou remained at 0.056–0.058 km2 for 50 years. Forest patches with areas smaller than the average accounted for more than 80% of the total number of patches.
The spatial distribution change of Picea crassifolia Kom. forest area from 1968 to 2017 is shown in Figure 5. Between 1968 and 2008, patch areas that remained unchanged accounted for 92.14% of the total area of the watershed. During that time, the ratio of Picea crassifolia Kom. area to non-Picea crassifolia Kom. area was 2.24%, while conversion of areas from non-Picea crassifolia Kom. to Picea crassifolia Kom. reached 5.62%.
Between 2008 and 2017, the area where the Picea crassifolia Kom. forest remained unchanged accounted for 97.49% of the total area of the basin. During that time, the proportion of Picea crassifolia Kom. area converted to non-Picea crassifolia Kom. was 0.89%, while the proportion of non-Picea crassifolia Kom. converted to Picea crassifolia Kom. area was 1.62%. Further, the most Picea crassifolia Kom. forest changes in area occurred at the boundaries of forest patches.
The rate of increase in patch area of the Picea crassifolia Kom. forest was 50.30 m2/per year from 1968 to 2008, and it decreased to 40.42 m2/per year from 2008 to 2017.
The regulation of elevation and aspect of the distribution of Picea crassifolia Kom. forest is shown in Figure 6. The altitude range of forest distribution in 2017 was 2639.70 to 3412.80 m, with the mean elevation of the distribution of 266 forest patches ranging from 2674.49 to 3261.26 m. The altitude of the lower tree line ranged from 2639.66 to 3220.54 m, and of the upper tree line from 2681.18 to 3412.77 m.
The average elevation of forest distribution increased from 2994.59 to 3023.90 m from 1968 to 2008, with a rate of increase of 2.44 m/per year, but it declined by 1.34 m and a decline rate of 0.50 m/per year by 2017.
The average altitude of the lower tree line increased by 23.74 m between 1968 and 2008, with a rate of 2.02 m/per year. However, this rate decreased by 1.20 m/per year during 2008–2017.
Throughout the study period, the upper tree line moved upward. During the period 1968–2008, the tree line moved 39.10 m with a rate of 3.12 m/per year. The average rate of the upward movement from 2008–2017 was 0.04 m/per year.
The forest was distributed over a wide range of slopes from 0 to 83.99°. The average slope of forest distribution was about 33°, and slope of the upper tree line, lower tree line, and mean forest did not change significantly. The Picea crassifolia Kom. forest was mainly distributed on shaded and semi-shaded slopes, and those accounted for more than 80% of the forest area, reaching 86.86% in 1968. From 1968 to 2017, the proportion of forests distributed on shaded slopes decreased, and the proportion of forests distributed on partly sunny slopes increased significantly.
The elevation distribution index over the past 50 years (Figure 7A) showed Picea crassifolia Kom. forests dominated elevations from 2700.00 to 3250.00 m for 50 years. Forest area increased from 1968 to 2008, and the increase was greatest at 2800.00–2900.00 m.
Changes in slope distribution index indicated that Picea crassifolia Kom. forest in the past 50 years dominated 10–35° and around 60° slopes (Figure 7B). The forest area on relatively flat slopes increased rapidly from 1968 to 2008. Further, Picea crassifolia Kom. forest was mainly found on shaded and semi-shaded slopes; however, the proportion of forest area on shaded slopes decreased between 1968 and 2008, while that on semi-sunny slopes increased (Figure 7C).

3.2. Landscape Changes

The largest patch index (LPI) did not exhibit a clear response function with the change in scale, and there was a large fluctuation at 2, 4, and 5 m (Figure 8). However, beyond 5 m, the exponential change was steady. PAFRAC showed an increasing trend with an increase in scale, and its landscape index scale maps fluctuated slightly at 2, 8, and 10 m. AI showed a downward trend. PD increased between 1 and 6 m and then decreased, with inflection points at 10 and 18 m. Although SPLIT decreased with increasing scale, no apparent relationship was detected, and the trend was opposite of that of LPI. Large fluctuations with a downward trend were found at 2, 4, and 5 m in the SPLIT map, and after 5 m, the exponential change stabilized. AI and LSI decreased linearly with inflection points at 3 and 10 m for AI, and at 4 and 12 m for LSI. Based on the calculated scale effects of the landscape index, the first scale domain of each landscape index and the appropriate scale for research were determined (Table 2). Finally, we found that the most appropriate landscape scale for this study was 7 m.
We used that scale to calculate changes in the six landscape pattern indices over the past 50 years (Figure 9).
The LPI were <30 in the past 50 years, and <20 in 1968, indicating that scattered rather than concentrated patches dominated in the landscape. After 1968, the index increased, but the value remained low, indicating that the development of Picea crassifolia Kom. was mainly due to the development of small patches. The calculated value of PAFRAC at about 1.32–1.34 in the past 50 years showed that the shape of forest patches was relatively regular, while the shape of the boundary tended to change from regular to irregular. In the past 50 years, forest PD, that is, patch uniformity, has been decreasing, but the change was relatively small. During last 50 years, SPLIT for the forest decreased, especially from 1968 to 2008, forest patches have developed rapidly, and patches became closer and some even merged. Aggregation Index (AI) did not change greatly in the past 50 years, with a difference from 1968 to 2017 of less than 0.1, indicating that the degree of aggregation did not change. From LSI, patch concentration in the Picea crassifolia Kom. forest has increased in the past 50 years, and the distance between patches became smaller, indicating that patches have expanded to a certain extent.

4. Discussion

We used high-resolution remote sensing images for three periods in this study to interpret the distribution of Picea crassifolia Kom. forest. Then, we discerned changes in the position of the tree line and attributed them to elevation, slope, aspect, and related indices. We determined characteristics and dynamic changes of the landscape pattern of Picea crassifolia Kom. forest with the landscape pattern index. Further, we also analyzed possible reasons for Picea crassifolia Kom. forest changes.

4.1. Landscape Pattern Characteristics of Picea crassifolia Kom. Forest

We found that Picea crassifolia Kom. forest in the Dayekou watershed was mostly distributed in patches and strips. Patchy forests become typical landscape types in arid and semi-arid areas as a result of interactions among accumulation and redistribution of water, nutrients, and other resources, and specific micro-topography [51,52]. High-resolution remote sensing images can monitor this type of vegetation better than other means [53]. The area and aggregation degree of Picea crassifolia Kom. forest patches were small. Small patch habitats had more conservation importance than large and medium forest patches [54], but small patches with short distances among them are susceptible to natural disasters and the ecosystem is relatively fragile [55].
The distribution regulation of the spruce forest in Qinghai is more obvious in elevation and aspect than slope. The range of altitudinal distribution of Picea crassifolia Kom. forest in Dayekou watershed of 2640 to 3413 m was consistent with earlier reports of the Picea crassifolia Kom. range of 2600 to 3400 m in the Qilian Mountains [56]. Our study also showed that Picea crassifolia Kom. forest in this watershed dominated at elevations of 2700 to 3250 m for 50 years, which is somewhat higher than the 2650–3100 m observed for the Qilian Mountains [56]. An analysis of stomatal density and its distribution pattern on epidermis, of the length and dry weight of needles, and of tree-ring, showed that altitude of about 3000 m was the optimum zone for growth and development of Picea crassifolia Kom. [57]. The apparent preference of Picea crassifolia Kom. forests for growth in this watershed on shaded slopes that are either relatively flat and shaded and half-shaded slopes is consistent with forest landscape patterns observed in the Qilian Mountains [58] and are apparently strongly related to humidity and solar radiation [30].

4.2. Landscape Pattern Change in Dayekou Catchment

More than 90% of the landscape area in the watershed did not change. However, the number of patches and patch area of the Picea crassifolia Kom. forest in Dayekou have been increasing since 1968, especially on flat slopes and partly sunny slope aspects. Rising tree lines have been observed in Europe, North America, and New Zealand [59,60], and at high latitudes and altitudes, 52% of mountain tree lines have shifted upward [61]. The expansion of mountain forests to higher altitudes and latitudes, and the upward movement of vegetation, are also frequently observed in high mountains [62,63]. In our study, forest interior was also vegetated during the study period, especially between 1968 and 2008.
The apparent expansion and upward movement of Picea crassifolia Kom. forest in Dayekou differed from the Picea crassifolia Kom. forest in Tianshan, northwestern China, where forest change was mainly reflected in an accelerated growth rate, while the tree line was not affected [64]. In Qilian Mountains, NDVI of vegetation appeared to have been increasing during 1982–2014 [65] and the carbon mass of Picea crassifolia Kom. forest in the Qilian Mountains has increased by 1.202 kg/m2 from 1964 to 2013 [66]. One survey showed that Picea crassifolia Kom. forest population density increased by 23 times at the tree line in the Qilian Mountains, but the tree line position was not significantly altered during the past 100 years [67]. Another survey determined that the elevation of the upper tree line shifted upward between 6.1 to 10.4 m from 1957 to 1980 [16]. We observed in this study an increase in forest density, an expansion of forest area, and an upward trend in the tree line of Picea crassifolia Kom. forest in the Qilian Mountains. Research on the Qinghai spruce forests in the Qilian Mountains focus mainly on establishing the relationships between tree growth and climate change through tree-rings [68,69,70]. Articles discussing changes in Picea crassifolia Kom. forest area and landscape pattern are relatively few.

4.3. Possible Reasons of Picea crassifolia Kom. Forest to Changes

Historically, forests in the Qilian Mountains underwent deforestation in the 1960s and 1970s [71]. But then, total area of Picea crassifolia Kom. forest in Dayekou increased, reflecting the trend of increasing vegetation cover in the Qilian Mountains after 1982 [65], especially following the establishment of the Qilian Mountain Reserve and declaration of the Picea crassifolia Kom. forest as an important ecosystem for water conservation and management [72]. Since the 1970s, in some areas of Qilian Mountains, grasslands on semi-shaded and partly sunny slopes had been converted into Picea crassifolia Kom. plantation forests [73]. Thus, the growth in Picea crassifolia Kom. forest patches and on flat sunny slopes during the study period may be caused by plantation establishment.
In the past 50 years, the average annual temperature and annual precipitation increased in Picea crassifolia Kom. growing area in the Qilian Mountains [66]. The increase in temperature due to climate change is more significant in mountainous areas, resulting in expansion of forests [74]. Picea crassifolia Kom. growth at the upper tree line in the Qilian Mountains appears to have responded to warmer conditions, too. The slower rate of forest expansion or even a shrinkage after 2008 may be related to the slowing of the temperature rise after 2000 [75,76]. The elevation distribution of the upper tree line of Picea crassifolia Kom. forest is mainly related to temperature, while the distribution of the lower tree line is related to precipitation. The rising lower tree line may be due to an increase in evapotranspiration and drought events caused by the temperature increase offsetting the expansion of the lower tree line due to precipitation [77].

5. Conclusions

We applied remote sensing images with high resolution to detect the slowly changing Picea crassifolia Kom. tree line. We found that the scale of detected change in Picea crassifolia Kom. forest was about or slightly below 30 m, which is difficult to monitor with low-resolution remote sensing images, such as Landsat. Therefore, to detect slowly evolving changes in coniferous forests in high mountain areas, it is necessary to use high-resolution imagery.
The number and area of forest patches have been increasing from 1968 to 2017, especially at relatively flat and partly sunny areas. Also, forest patch concentration increased. The patches have expanded, and their upper and lower tree line and average altitude have moved upward. However, the rate of area increase after 2008, and the upward trend of forest lines, started to slow down. Recent changes in Picea crassifolia Kom. forests may be related to the establishment of forest plantations and recent climate change.
Because the selected remote sensing image had only three time periods, additional remote sensing data will be added to improve the correlation analysis between forest change and climate factors at a future time.

Author Contributions

Z.H. and S.F. conceived and designed the experiments; S.F. performed the experiments, analyzed the data and wrote the paper. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the PhD early development program of Shangluo University (19SKY027), the National Key Research and Development Program of China (No.2017YFC0504306), the Strategic Priority Research Program of the Chinese Academy of Sciences (No.Y92C782001), and the National Natural Science Foundation of China (No. 41901050).

Acknowledgments

We are very grateful to Kathryn Piatek for her comments and editorial assistance.

Conflicts of Interest

The authors declare no conflict of interest. The funding sponsors had no role in the design of the study, data collection, analyses, interpretation, writing of the manuscript, or in the decision to publish the results.

References

  1. Vanneste, T.; Michelsen, O.; Graae, B.J.; Kyrkjeeide, M.O.; Holien, H.; Hassel, K.; Lindmo, S.; Kapás, R.E.; Frenne, P.D. Impact of climate change on alpine vegetation of mountain summits in Norway. Ecol. Res. 2017, 32, 1–15. [Google Scholar] [CrossRef]
  2. Körner, C. A re-assessment of high elevation treeline positions and their explanation. Oecologia 1998, 115, 445–459. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  3. Wieser, G. Lessons from the timberline ecotone in the Central Tyrolean Alps: A review. Plant Ecol. Divers. 2012, 5, 127–139. [Google Scholar] [CrossRef]
  4. Hofgaard, A.; Dalen, L.; Hytteborn, H. Tree recruitment above the treeline and potential for climate-driven treeline change. J. Veg. Sci. 2009, 20, 1133–1144. [Google Scholar] [CrossRef]
  5. Malanson, G.P. Mixed signals in trends of variance in high-elevation tree ring chronologies. J. Mt. Sci. 2017, 14, 1961–1968. [Google Scholar] [CrossRef]
  6. Hagedorn, F.; Gavazov, K.; Alexander, J.M. Above-and belowground linkages shape responses of mountain vegetation to climate change. Science 2019, 365, 1119–1123. [Google Scholar] [CrossRef]
  7. Tasser, E.; Leitinger, G.; Tappeiner, U. Climate change versus land-use change—What affects the mountain landscapes more? Land Use Policy 2017, 60, 60–72. [Google Scholar] [CrossRef]
  8. Du, H.; Liu, J.; Li, M.H.; Büntgen, U.; Yang, Y.; Wang, L.; Wu, Z.; He, H.S. Warming-Induced upward migration of the alpine treeline in the Changbai Mountains, northeast China. Glob. Chang. Biol. 2018, 24, 1256–1266. [Google Scholar] [CrossRef]
  9. Shi, C.; Shen, M.; Wu, X.; Cheng, X.; Li, X.; Fan, T.; Li, Z.; Zhang, Y.; Fan, Z.; Shi, F. Growth response of alpine treeline forests to a warmer and drier climate on the southeastern Tibetan Plateau. Agric. For. Meteorol. 2019, 264, 73–79. [Google Scholar] [CrossRef]
  10. Franke, A.; Bräuning, A.; Timonen, M.; Rautio, P. Growth response of Scots pines in polar-alpine tree-line to a warming climate. For. Ecol. Manag. 2017, 399, 94–107. [Google Scholar] [CrossRef]
  11. Chang, X.; Zhao, W.; He, Z. Radial pattern of sap flow and response to microclimate and soil moisture in Qinghai spruce (Picea crassifolia) in the upper Heihe River Basin of arid northwestern China. Agric. For. Meteorol. 2014, 187, 14–21. [Google Scholar] [CrossRef]
  12. Chang, X.; Zhao, W.; Liu, H.; Wei, X.; Liu, B.; He, Z. Qinghai spruce (Picea crassifolia) forest transpiration and canopy conductance in the upper Heihe River Basin of arid northwestern China. Agric. For. Meteorol. 2014, 198, 209–220. [Google Scholar] [CrossRef]
  13. Tian, Q.; He, Z.; Xiao, S.; Peng, X.; Ding, A.; Lin, P. Response of stem radial growth of Qinghai spruce (Picea crassifolia) to environmental factors in the Qilian Mountains of China. Dendrochronologia 2017, 44, 76–83. [Google Scholar] [CrossRef]
  14. Lingnan, Z.; Shuangshuang, L.; Yixue, H.; Xiaomin, Z.; Xiaohong, L. Changes in the radial growth of Picea crassifolia and its driving factors in the mid-western Qilian Mountains, Northwest China since 1851 CE. Dendrochronologia 2020, 61, 125707. [Google Scholar]
  15. Liang, E.; Shao, X.; Eckstein, D.; Liu, X. Spatial variability of tree growth along a latitudinal transect in the Qilian Mountains, northeastern Tibetan Plateau. Can. J. For. Res. 2010, 40, 200–211. [Google Scholar] [CrossRef]
  16. He, Z.; Zhao, W.; Zhang, L.; Liu, H. Response of tree recruitment to climatic variability in the alpine treeline ecotone of the Qilian Mountains, northwestern China. For. Sci. 2013, 59, 118–126. [Google Scholar] [CrossRef]
  17. Morley, P.J.; Donoghue, D.N.; Chen, J.-C.; Jump, A.S. Quantifying structural diversity to better estimate change at mountain forest margins. Remote Sens. Environ. 2019, 223, 291–306. [Google Scholar] [CrossRef]
  18. Kohler, T. Mountains and Climate Change; Center for Development and Environment: Bern, Switzerland, 2015. [Google Scholar]
  19. Kurz, W.A.; Dymond, C.; Stinson, G.; Rampley, G.; Neilson, E.; Carroll, A.; Ebata, T.; Safranyik, L. Mountain pine beetle and forest carbon feedback to climate change. Nature 2008, 452, 987–990. [Google Scholar] [CrossRef]
  20. Holtmeier, F.-K.; Broll, G. Treeline research—From the roots of the past to present time. A review. Forests 2020, 11, 38. [Google Scholar] [CrossRef] [Green Version]
  21. Baker, W.L.; Honaker, J.J.; Weisberg, P.J. Using aerial photography and GIS to map the forest-tundra ecotone in Rocky Mountain National Park, Colorado, for global change research. Photogramm. Eng. Remote Sens. 1995, 61, 313–320. [Google Scholar]
  22. Panigrahy, S.; Anitha, D.; Kimothi, M.; Singh, S. Timberline change detection using topographic map and satellite imagery. Trop. Ecol. 2010, 51, 87–91. [Google Scholar]
  23. Olthof, I.; Pouliot, D. Treeline vegetation composition and change in Canada’s western Subarctic from AVHRR and canopy reflectance modeling. Remote Sens. Environ. 2010, 114, 805–815. [Google Scholar] [CrossRef]
  24. Xu, D.; Geng, Q.; Jin, C.; Xu, Z.; Xu, X. Tree line identification and dynamics under climate change in Wuyishan National Park based on Landsat images. Remote Sens. 2020, 12, 2890. [Google Scholar] [CrossRef]
  25. Chen, Y.; Lu, D.; Luo, G.; Huang, J. Detection of vegetation abundance change in the alpine tree line using multitemporal Landsat Thematic Mapper imagery. Int. J. Remote Sens. 2015, 36, 4683–4701. [Google Scholar] [CrossRef]
  26. Bolton, D.K.; Coops, N.C.; Hermosilla, T.; Wulder, M.A.; White, J.C. Evidence of vegetation greening at alpine treeline ecotones: Three decades of Landsat spectral trends informed by lidar-derived vertical structure. Environ. Res. Lett. 2018, 13, 084022. [Google Scholar] [CrossRef]
  27. Franke, A.; Feilhauer, H.; Bräuning, A.; Rautio, P.; Braun, M. Remotely sensed estimation of vegetation shifts in the polar and alpine tree-line ecotone in Finnish Lapland during the last three decades. For. Ecol. Manag. 2019, 454, 117668. [Google Scholar] [CrossRef]
  28. Mathisen, I.E.; Mikheeva, A.; Tutubalina, O.V.; Aune, S.; Hofgaard, A. Fifty years of tree line change in the Khibiny Mountains, Russia: Advantages of combined remote sensing and dendroecological approaches. Appl. Veg. Sci. 2013, 17, 6–16. [Google Scholar] [CrossRef] [Green Version]
  29. Wang, Q.; Wu, N.; Luo, P.; Yi, S.; Bao, W.; Shi, F. Growth rate of mosses and their environmental determinants in subalpine coniferous forests and clear-cuts at the eastern edge of the Qinghai-Tibetan Plateau, China. Front. For. China 2008, 3, 171–176. [Google Scholar] [CrossRef] [Green Version]
  30. He, Z.-B.; Fang, S.; Chen, L.-F.; Du, J.; Zhu, X.; Lin, P.-F. Spatial patterns in natural Picea crassifolia forests of northwestern China, as basis for close-to-nature forestry. J. Mt. Sci. 2018, 15, 1909–1919. [Google Scholar] [CrossRef]
  31. Fowler, M.J.F. Archaeology through the keyhole: The serendipity effect of aerial reconnaissance revisited. Interdiscip. Sci. Rev. 2004, 29, 118–134. [Google Scholar]
  32. Yang, C.; Everitt, J.H.; Bradford, J.M. Comparison of QuickBird satellite imagery and Airborne imagery for mapping Grain Sorghum Yield patterns. Precis. Agric. 2006, 7, 33–44. [Google Scholar] [CrossRef]
  33. Wang, T.; Zhang, G.; Li, D.; Tang, X.; Jiang, Y.; Pan, H.; Zhu, X.; Fang, C. Geometric accuracy validation for ZY-3 satellite imagery. IEEE Geoence Remote Sens. Lett. 2014, 11, 1168–1171. [Google Scholar] [CrossRef]
  34. Jing, W. WATER: Dataset of Forest Structure Parameter Survey at the Fixed Sampling Plot in the Pailugou Watershed and Dayekou Watershed Foci Experiment Area in 2003; Academy of Water Resources Convervation Forests in Qilian Mountain of Gansu Province: Lanzhou, China, 2012. [Google Scholar] [CrossRef]
  35. Ma, G.; Jing, W. WATER: Dataset of Forest Structure Parameter Survey at the Fixed Sampling Plot in the Pailugou Watershed and Dayekou Watershed Foci Experiment Area in 2007; Academy of Water Resources Convervation Forests in Qilian Mountain of Gansu Province: Lanzhou, China, 2013. [Google Scholar] [CrossRef]
  36. Chen, X.; Guo, Z.; Jin, M. WATER: Dataset of Forest Structure Parameter Survey at the Temporary Forest Sampling Plot in the Dayekou Watershed Foci Experimental Area; Institute of Forest Resource Information Techniques, Chinese Academy of Forestry, Institute of Remote Sensing Applications: Beijing, China; Chinese Academy of Sciences, Academy of Water Resources Conservation Forests in Qilian Mountain of Gansu Province: Beijing, China, 2008. [Google Scholar] [CrossRef]
  37. Bai, Y.; Liu, Z.; Fu, Z.; Li, B.; Lin, H.; Song, D.; Sun, Z.; Gong, H.; Zhu, M. WATER: Dataset of Land Use and Land Cover Investigation in the Arid Region Hydrology and Forest Hydrology Experiment Areas; National Tibetan Plateau Data Center: Beijing, China, 2008. [Google Scholar] [CrossRef]
  38. Song, J.; Fu, Z.; Zhang, X.; Wang, S.; Zhao, M.; Lei, J.; Niu, Y.; Lun, L.; Liang, F.; He, Q. WATER: Dataset of Forest Structure Parameter Measurements for the Fixed Forest Sampling Plots in the Dayekou and Pailugou Watershed Foci Experimental Areas (2003–3007); National Tibetan Plateau Data Center: Beijing, China, 2008. [Google Scholar] [CrossRef]
  39. Zhang, Y.; Ma, M. HiWATER: 1m DEM Data Production in Dayekou Watershed; National Tibetan Plateau Data Center: Beijing, China, 2014. [Google Scholar] [CrossRef]
  40. Guide, E.U.S. ENVI On-Line Software User’s Manual; ITT Visual Information Solutions: Boulder, CO, USA, 2008. [Google Scholar]
  41. Lassiter, A.; Darbari, M. Assessing alternative methods for unsupervised segmentation of urban vegetation in very high-resolution multispectral aerial imagery. PLoS ONE 2020, 15, e0230856. [Google Scholar] [CrossRef]
  42. ESRI. (Environmental Sciences Research Institute). 2012. Available online: http://www.esri.com/ (accessed on 11 June 2012).
  43. Wardrop, D.; Bishop, J.; Easterling, M.; Hychka, K.; Myers, W.; Patil, G.; Taillie, C. Use of landscape and land use parameters for classification and characterization of watersheds in the mid-Atlantic across five physiographic provinces. Environ. Ecol. Stat. 2005, 12, 209–223. [Google Scholar] [CrossRef]
  44. Peng, J.; Wang, Y.; Zhang, Y.; Wu, J.; Li, W.; Li, Y. Evaluating the effectiveness of landscape metrics in quantifying spatial patterns. Ecol. Indic. 2010, 10, 217–223. [Google Scholar] [CrossRef]
  45. Fang, S.; Zhao, Y.; Han, L.; Ma, C. Analysis of landscape patterns of arid valleys in China, based on grain size effect. Sustainability 2017, 9, 2263. [Google Scholar] [CrossRef] [Green Version]
  46. Mcgarigal, K.S.; Cushman, S.A.; Neel, M.C.; Ene, E. FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. 2002. Available online: www.umass.edu/landeco/research/fragstats/fragstats.htmlS (accessed on 2 November 2017).
  47. Tarantino, C.; Adamo, M.; Lucas, R.; Blonda, P. Detection of changes in semi-natural grasslands by cross correlation analysis with WorldView-2 images and new Landsat 8 data. Remote Sens. Environ. 2016, 175, 65–72. [Google Scholar] [CrossRef]
  48. Roy, D.P.; Wulder, M.; Loveland, T.R.; Woodcock, C.; Allen, R.; Anderson, M.; Helder, D.; Irons, J.; Johnson, D.; Kennedy, R. Landsat-8: Science and product vision for terrestrial global change research. Remote Sens. Environ. 2014, 145, 154–172. [Google Scholar] [CrossRef] [Green Version]
  49. Foody, G.M. Classification accuracy comparison: Hypothesis tests and the use of confidence intervals in evaluations of difference, equivalence and non-inferiority. Remote Sens. Environ. 2009, 113, 1658–1663. [Google Scholar] [CrossRef] [Green Version]
  50. Congalton, R.G. Accuracy assessment and validation of remotely sensed and other spatial information. Int. J. Wildland Fire 2001, 10, 321–328. [Google Scholar] [CrossRef] [Green Version]
  51. Du, J.; Yan, P.; Dong, Y. Water driving mechanism of patched vegetation formation in arid areas: A review. Chin. J. Ecol. 2012, 31, 2137–2144, (In Chinese with English abstract). [Google Scholar]
  52. Assal, T.J.; Anderson, P.J.; Sibold, J. Spatial and temporal trends of drought effects in a heterogeneous semi-arid forest ecosystem. For. Ecol. Manag. 2016, 365, 137–151. [Google Scholar] [CrossRef] [Green Version]
  53. Liu, Q.; Zhang, Y.; Liu, G.; Huang, C.; Chai, S. Comparison of different spatial resolution images of ZY-3 satellite to patch vegetation detection. Bull. Surv. Mapp. 2014, 46, 16–20, (In Chinese with English abstract). [Google Scholar]
  54. Erdős, L.; Kröel-Dulay, G.; Bátori, Z.; Kovács, B.; Németh, C.; Kiss, P.J.; Tölgyesi, C. Habitat heterogeneity as a key to high conservation value in forest-grassland mosaics. Biol. Conserv. 2018, 226, 72–80. [Google Scholar] [CrossRef] [Green Version]
  55. Yang, G.; Xiao, D.; Zhao, C. GIS-Based analysis on the forest landscape patterns in the Qil ian Mountain. Arid Zone Res. 2004, 21, 27–32, (In Chinese with English abstract). [Google Scholar]
  56. Zhao, C.; Nan, Z.; Cheng, G.; Zhang, J.; Feng, Z. GIS-Assisted modelling of the spatial distribution of Qinghai spruce (Picea crassifolia) in the Qilian Mountains, northwestern China based on biophysical parameters. Ecol. Model. 2006, 191, 487–500. [Google Scholar] [CrossRef]
  57. Qiang, W.Y.; Wang, X.L.; Chen, T.; Feng, H.Y.; An, L.Z.; He, Y.Q.; Wang, G. Variations of stomatal density and carbon isotope values of Picea crassifolia at different altitudes in the Qilian Mountains. Trees 2003, 17, 258–262. [Google Scholar] [CrossRef]
  58. Guojing, Y.; Duning, X.; Lihua, Z.; Cuiwen, T. Hydrological effects of forest landscape patterns in the Qilian Mountains. Mt. Res. Dev. 2005, 25, 262–268. [Google Scholar] [CrossRef] [Green Version]
  59. Parmesan, C.; Yohe, G. A globally coherent fingerprint of climate change impacts across natural systems. Nature 2003, 421, 37–42. [Google Scholar] [CrossRef]
  60. Pauli, H.; Gottfried, M.; Reiter, K.; Klettner, C.; Grabherr, G. Signals of range expansions and contractions of vascular plants in the high Alps: Observations (1994–2004) at the GLORIA* master site Schrankogel, Tyrol, Austria. Glob. Chang. Biol. 2007, 13, 147–156. [Google Scholar] [CrossRef]
  61. Harsch, M.A.; Hulme, P.E.; Mcglone, M.S.; Duncan, R.P. Are treelines advancing? A global meta-analysis of treeline response to climate warming. Ecol. Lett. 2009, 12, 1040. [Google Scholar] [CrossRef] [PubMed]
  62. MacDonald, G.; Kremenetski, K.; Beilman, D. Climate change and the northern Russian treeline zone. Philos. Trans. R. Soc. Lond. B Biol. Sci. 2008, 363, 2283–2299. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  63. Wilson, R.J.; Gutiérrez, D.; Gutiérrez, J.; Martínez, D.; Agudo, R.; Monserrat, V.J. Changes to the elevational limits and extent of species ranges associated with climate change. Ecol. Lett. 2005, 8, 1138–1146. [Google Scholar] [CrossRef] [PubMed]
  64. Qi, Z.; Liu, H.; Wu, X.; Hao, Q. Climate-Driven speedup of alpine treeline forest growth in the Tianshan Mountains, Northwestern China. Glob. Chang. Biol. 2015, 21, 816–826. [Google Scholar] [CrossRef] [PubMed]
  65. Jia, W.; Chen, J. Variations of NDVI and its response to climate change in the growing season of vegetation in Qilianshan Mountains from 1982 to 2014. Res. Soil Water Conserv. 2018, 25, 264–268, (In Chinese with English abstract). [Google Scholar]
  66. Fang, S.; He, Z.; Du, J.; Chen, L.; Lin, P.; Zhao, M. Carbon mass change and its drivers in a Boreal coniferous forest in the Qilian Mountains, China from 1964 to 2013. Forests 2018, 9, 57. [Google Scholar] [CrossRef] [Green Version]
  67. Zhang, L.; Liu, H. Response of Picea crassifolia population to climate change at the treeline ecotones in Qilian Mountains. Sci. Silvae Sin. 2012, 48, 18–21, (In Chinese with English abstract). [Google Scholar]
  68. Wang, B.; Chen, T.; Li, C.; Xu, G.; Wu, G.; Liu, G. Radial growth of Qinghai spruce (Picea crassifolia Kom.) and its leading influencing climate factor varied along a moisture gradient. For. Ecol. Manag. 2020, 476, 118474. [Google Scholar] [CrossRef]
  69. Liang, E.; Leuschner, C.; Dulamsuren, C.; Wagner, B.; Hauck, M. Global warming-related tree growth decline and mortality on the north-eastern Tibetan plateau. Clim. Chang. 2016, 134, 163–176. [Google Scholar] [CrossRef]
  70. Gou, X.; Chen, F.; Yang, M.; Li, J.; Peng, J.; Jin, L. Climatic response of thick leaf spruce (Picea crassifolia) tree-ring width at different elevations over Qilian Mountains, northwestern China. J. Arid Environ. 2005, 61, 513–524. [Google Scholar] [CrossRef]
  71. Liu, Z.; Zhao, C.; Bai, Y.; Peng, S.; Nan, Z.; Liu, X.; Hao, H. Difference in stem volume of Qinghai spruce (Picea crassifolia) in catchments of Qilian Mountains. J. Lanzhou Univ. Nat. Sci. 2013, 49, 747–751, (In Chinese with English abstract). [Google Scholar]
  72. Zhao, C.; Bie, Q.; Peng, H. Analysis of the niche space of Picea crassifolia on the northern slope of Qilian Mountains. Acta Geogr. Sin. 2010, 65, 113–121, (In Chinese with English abstract). [Google Scholar]
  73. Zhu, X.; He, Z.; Chen, L.; Du, J.; Yang, J.; Lin, P.; Li, J. Changes in species diversity, aboveground biomass, and distribution characteristics along an afforestation successional gradient in semiarid Picea crassifolia plantations of Northwestern China. For. Sci. 2016, 63. [Google Scholar] [CrossRef]
  74. De Wit, H.A.; Bryn, A.; Hofgaard, A.; Karstensen, J.; Peters, G.P. Climate warming feedback from mountain birch forest expansion: Reduced albedo dominates carbon uptake. Glob. Chang. Biol. 2014, 20, 2344. [Google Scholar] [CrossRef] [PubMed]
  75. Wang, X.; Chen, R.; Liu, J. Spatial and temporal variation characteristics of accumulated negative temperature in Qilian Mountains under climate change. Plateau Meteorol. 2017, 36, 1267–1275, (In Chinese with English abstract). [Google Scholar]
  76. Cao, G.; Fu, J.; Li, L.; Cao, S.; Tang, Z.; Jiang, G.; Yu, M.; Yuan, J.; Han, G.; Diao, E. Analysis on temporal and spatial variation characteristics of air temperature in the south slope of Qilian Mountains and its nearby regions during the period from 1960 to 2014. Res. Soil Water Conserv. 2018, 25, 88–96, (In Chinese with English abstract). [Google Scholar]
  77. Yang, W.; Wang, Y.; Web, A.A.; Li, Z.; Tian, X.; Han, Z.; Wang, S.; Yu, P. Influence of climatic and geographic factors on the spatial distribution of Qinghai spruce forests in the dryland Qilian Mountains of Northwest China. Sci. Total Environ. 2018, 612, 1007. [Google Scholar] [CrossRef]
Forests 11 01188 g001 550
Figure 1. Location of Dayekou watershed in the Qilian Mountains of northwestern China.
Figure 1. Location of Dayekou watershed in the Qilian Mountains of northwestern China.
Forests 11 01188 g001
Forests 11 01188 g002 550
Figure 2. High-resolution digital elevation model (DEM) and verification points.
Figure 2. High-resolution digital elevation model (DEM) and verification points.
Forests 11 01188 g002
Forests 11 01188 g003 550
Figure 3. Verification data for drone sampling sites.
Figure 3. Verification data for drone sampling sites.
Forests 11 01188 g003
Forests 11 01188 g004 550
Figure 4. Distribution of Picea crassifolia Kom. in the Dayekou watershed in different years.
Figure 4. Distribution of Picea crassifolia Kom. in the Dayekou watershed in different years.
Forests 11 01188 g004
Forests 11 01188 g005 550
Figure 5. Area change of Picea crassifolia Kom. from 1968 to 2008 and 2008 to 2017 in the Dayekou watershed.
Figure 5. Area change of Picea crassifolia Kom. from 1968 to 2008 and 2008 to 2017 in the Dayekou watershed.
Forests 11 01188 g005
Forests 11 01188 g006 550
Figure 6. Elevation, slope, and aspect features of Picea crassifolia Kom. forest distribution.
Figure 6. Elevation, slope, and aspect features of Picea crassifolia Kom. forest distribution.
Forests 11 01188 g006
Forests 11 01188 g007 550
Figure 7. Change in the (A) elevation, (B) slope, and (C) aspect distribution index (y-axis is the value calculated with Equation (1), and x-axis is the levels of elevation, slope, and aspect).
Figure 7. Change in the (A) elevation, (B) slope, and (C) aspect distribution index (y-axis is the value calculated with Equation (1), and x-axis is the levels of elevation, slope, and aspect).
Forests 11 01188 g007
Forests 11 01188 g008 550
Figure 8. Response curve of landscape indices.
Figure 8. Response curve of landscape indices.
Forests 11 01188 g008
Forests 11 01188 g009 550
Figure 9. The change of landscape pattern.
Figure 9. The change of landscape pattern.
Forests 11 01188 g009
Table
Table 1. Landscape pattern indices.
Table 1. Landscape pattern indices.
MetricsNameDescription
Area and EdgeLargest Patch Index (LPI)The proportion of the largest patch area
ShapePerimeter-Area Fractal Dimension (PAFRAC)Non-randomness or degree of aggregation for different patches
AggregationPatch Density (PD)Number of patches per unit area
Splitting Index (SPLIT)The number of patches in a landscape divided into equal sizes keeping landscape division constant, express the separation degree of individual distribution in different
Aggregation Index (AI)The degree of aggregation of similar patches
Landscape Shape Index (LSI)Continuity and complexity of landscape shape and the measurement of the perimeter-to-area ratio for the landscape as a whole
Table
Table 2. The appropriate scale of the landscape metrics.
Table 2. The appropriate scale of the landscape metrics.
MetricsFirst Scale DomainThe Appropriate Scale
LPI2–5 m>5 m
PAFRAC2–8 m3–7 m
PD6–10 m7–9 m
SPLIT2–5 m>5 m
AI3–10 m2–9 m
LSI4–12 m3–11 m
All6–8 m7 m
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Fang, S.; He, Z. Fifty Years of Change in a Coniferous Forest in the Qilian Mountains, China—Advantages of High-Definition Remote Sensing. Forests 2020, 11, 1188. https://doi.org/10.3390/f11111188

AMA Style

Fang S, He Z. Fifty Years of Change in a Coniferous Forest in the Qilian Mountains, China—Advantages of High-Definition Remote Sensing. Forests. 2020; 11(11):1188. https://doi.org/10.3390/f11111188

Chicago/Turabian Style

Fang, Shu, and Zhibin He. 2020. "Fifty Years of Change in a Coniferous Forest in the Qilian Mountains, China—Advantages of High-Definition Remote Sensing" Forests 11, no. 11: 1188. https://doi.org/10.3390/f11111188

APA Style

Fang, S., & He, Z. (2020). Fifty Years of Change in a Coniferous Forest in the Qilian Mountains, China—Advantages of High-Definition Remote Sensing. Forests, 11(11), 1188. https://doi.org/10.3390/f11111188

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop