Qinghai Lake Basin area in Gangcha county is selected as the study area in terms of desertification change features in this paper. Based on the remote sensing (RS) and global positioning system (GPS) technologies, the desertification information range from 1989 to 2014 in the study area is extracted. Using the method of the decision tree, the desertification in the research area is been divided into four grades including mild desertification, moderate desertification, severe desertification and serious desertification. The change characteristics of desertification in the study area were analyzed in detail, which showed that the desertification in the study area experienced a process of first development and then a reversal. The rapid development of desertification appears in the 1990s, where about 1101.22 kilometers of desertification area was increased in this stage. Since the twenty-first Century, the desertification is gradually significant recovered and local area exist intensified desertification. There are tendencies of interactive transform in different types of desertification. The tendencies of different degrees of desertification land are rising, and there’re some differences in rising rates, where the expansion rate of moderate desertification is the biggest, increasing by 7.27 kilometers per year.
Desertification is a manifestation of land degradation, which was first proposed by Lavauden [
Desertification mapping is the important and fundamental work for studies of the mechanism of desertification, spatial and temporal change of land degradation. Hanan et al. [
The sensitivity of desertification in the surrounding area of the Qinghai Lake is moderate and high. In recent years, the desertification situation in the Qinghai Lake area has been quite serious. Therefore, the Qinghai Lake area has become one of the hot spots and focus areas of desertification around the world. Therefore, Gangcha County was selected as a representative county to carry out relevant research and provide scientific decision-making basis for desertification prevention and control.
The remainder of this paper is organized as follows. Section 2 describes the study area. Section 3 describes the data preprocessing. Then, Section 4 describes the research methods. Next, Section 5 demonstrates and discusses the research results in detail, and finally, Section 6 presents the conclusions.
Study area, on the north bank of Qinghai Lake, is located at Gangcha County, Haibei Tibetan Autonomous Prefecture, Qinghai Province. The geographical location is between 99°20 ′44 ″–100°37′ 24″ E, 36°58 ′06 ″–38°04′ 04″ N, bordering Haiyan County on the east, Tianjun County of Haixi Mongolian and Tibetan Autonomous Prefecture on the west, Buha River and Hainan County on the south, and Tonghe River on the north. The total area is 1.2*104 km2, and the land type is mainly grassland. The terrain of the study area is clear north high and south low, with an average altitude of more than 3300 m. Gangcha county is a typical highland continental climate, long sunshine time, the large temperature difference between day and night; The average annual precipitation is 370.5 mm, and the annual evaporation is 1500.6~1847.8 mm. It is cool in summer and autumn and cold in winter. The average annual temperature is only −0.6°C. In this paper, the southern lake basin area with obvious desertification development was selected as the research area (As shown in
This paper mainly uses Landsat 5 Thematic Mapper (TM) remote sensing image data to interpret and obtain the land cover in the study area. Landsat data is obtained by NASA, which is favored for its advantages of multi-band, good spatial resolution (30 M) and short monitoring period, especially its superior potential in long-term continuous remote sensing monitoring at the large regional scale. In this study, Landsat 5 TM images from 1989 to 2014 were selected as the main data source. The images were mainly taken from July to September when the vegetation coverage of the Gangcha Lake basin area was better and the thermal radiation difference of ground objects was more obvious, and the cloud content of the images was less than 10%. Specific image information is shown in
Serial number | Type of data | Imaging time | Track number | Number of bands | Spatial resolution (m) |
---|---|---|---|---|---|
1 | Landsat5 TM | 1989.08 | p133r34 | 7 | 30.0 |
2 | Landsat5 TM | 1990.06 | p133r34 | 7 | 30.0 |
3 | Landsat5 TM | 1993.08 | p133r34 | 7 | 28.5 |
4 | Landsat5 TM | 1995.08 | p133r34 | 7 | 30.0 |
5 | Landsat5 TM | 1996.06 | p133r34 | 7 | 28.5 |
6 | Landsat5 TM | 1997.08 | p133r34 | 7 | 28.5 |
7 | Landsat5 TM | 1998.07 | p133r34 | 7 | 28.5 |
8 | Landsat5 TM | 1999.07 | p133r34 | 7 | 28.5 |
9 | Landsat5 TM | 2000.08 | p133r34 | 7 | 30.0 |
10 | Landsat5 TM | 2001.07 | p133r34 | 7 | 30.0 |
11 | Landsat5 TM | 2002.07 | p133r34 | 7 | 30.0 |
12 | Landsat5 TM | 2003.09 | p133r34 | 7 | 30.0 |
13 | Landsat5 TM | 2004.09 | p133r34 | 7 | 30.0 |
14 | Landsat5 TM | 2005.09 | p133r34 | 7 | 30.0 |
15 | Landsat5 TM | 2006.08 | p133r34 | 7 | 30.0 |
16 | Landsat5 TM | 2008.07 | p133r34 | 7 | 30.0 |
17 | Landsat5 TM | 2009.08 | p133r34 | 7 | 30.0 |
18 | Landsat5 TM | 2010.07 | p133r34 | 7 | 30.0 |
19 | Landsat5 TM | 2011.06 | p133r34 | 7 | 30.0 |
20 | Landsat8 OLI | 2013.09 | p133r34 | 11 | 30.0 |
21 | Landsat8 OLI | 2014.07 | p133r34 | 11 | 30.0 |
Due to attenuation and satellite flight attitude sensor function, the effect of many factors, such as atmosphere and at the time of complex surface information, will inevitably produce a variety of radiation distortion, geometric distortion and atmospheric extinction system error and random error, which is to reduce the quality of the remote sensing image and affects the accuracy of image analysis [
On the basis of the interpretation of the desertification results, the desertification vector diagrams of 1990 and 2000, 2000 and 2009, and 2009 and 2014 were superimposed and calculated, obtained the years 1990 to 2000, 2000 to 2009, and the transfer matrix of various types of desertified land in the three periods, and analyzed the evolution mechanism of different types of desertified land in two different time dimensions: interdecadal and interannual.
Albedo refers to the ratio of solar radiation reflected by the earth’s surface, that is, the ratio of the solar radiation flux reflected by the Earth’s surface to the solar radiation flux incident. Surface Albedo, or surface Albedo, reflects the earth’s ability to reflect solar radiation. Radiation from the sun drives the land, sea, air the material exchange and energy cycle of the ecosystem, absorbed by the surface solar radiation conditions but also will affect the entire earth’s weather and climate change. So the surface albedo is the key factor of the development and changes of the earth’s climate system, it is an important parameter of the terrestrial surface radiation energy balance. Referring to the calculation method, this study uses the reflectance in the direction of the top of the atmosphere to estimate the surface albedo of the wideband, and the calculation formula is as follows:
Fractional Vegetation Cover (FVC) refers to the vertical projection area of Vegetation in the unit area [
According to the binary pixel model [
In the formula,
To reduce the influence of soil, shadow, atmospheric environment and other factors on the vegetation index inversion results, the improved soil-adjusted vegetation index (MSAVI) came into being. MSAVI can minimize the influence of soil factors and enhance sensitivity to vegetation [
Land Surface Temperature (LST), which is an important parameter in desertification monitoring, will change with the change of desertified Land types. Land surface temperature inversion is mainly based on remote sensing data with the thermal infrared band. At present, LST inversion algorithms mainly include atmospheric correction method, single-channel algorithm, split window algorithm, and multi-band algorithm. In this study, the single-channel algorithm proposed by Weng Qihao et al. was used to invert land surface temperature.
First, the gray value is converted into the emitted radiation energy of the object, and the formula is as follows:
Secondly, the radiation energy emitted by the object is converted into the brightness temperature of surface radiation, and the transformation formula is as follows:
Then, the brightness temperature can be converted to the surface temperature, and the calculation formula is as follows:
Where, λ = 11.5 um; H is Planck constant, 6.26 × 10−34, unit is J·S; C is the speed of light, and the value is 2.998 × 108 m/s. α is Stefan Boltzmann constant, with a value of 1.38 × 10−23, and the unit is J/K. ε is the specific emissivity. The modified specific emissivity method proposed by Qin [
In the formula,
Soil moisture (WET) is an important indicator for monitoring land degradation and is of great significance for remote sensing monitoring and quantitative evaluation of desertification. The vegetation canopy temperature is affected by soil moisture. Sandholt et al. [
Remote sensing image classification is the prerequisite to evaluate the degree and spatial distribution of land desertification in the study area. The nature of decision tree classification is a kind of supervised classification, and its classification steps can be roughly divided into the following steps: First, the definition of knowledge (rules), which can be expressed by mathematical language such as algorithm, or defined by experience summary; Second is the input of rules, that is, the defined classification rules are input to the classifier; Thirdly the decision tree operation is carried out; Finally, the classification results are processed after output. Before decision tree classification, it is necessary to fuse the inverted vegetation coverage, albedo, land surface temperature, improved soil-adjusted vegetation index, and soil moisture image to obtain a new multi-band remote sensing image, which is used as the input layer of decision tree classification. In order to refine the desertification grade of the study area, the desertification grade of the study area was divided into seven categories: mild desertification, moderate desertification, severe desertification, severe desertification, non-desertification land, water body, and others (including cloud, cloud shadow, and snow).
According to the mathematical statistics of inversion parameters, the classification thresholds of desertification degree of different grades were set. In general, NDVI value can distinguish the vegetation covered area from the non-vegetation covered area, and the vegetation coverage can effectively distinguish the mild desertification land, moderate desertification land, severe desertification land, and desertified land. Albedo can efficiently cloud, the cloud and snow from vegetation types in the land to the isolated, the surface temperature can distinguish between water and serious desertification land types, the modified soil adjusted vegetation index and soil moisture can further distinguish between vegetation and vegetation coverage less desertification, water, and different land types such as cloud. The specific classification rules of the decision tree are shown in
Established based on the decision tree classification system of land desertification in the study area can be divided into different levels, classification of inevitably in the salt and pepper effect, this is caused by noise classification error, the classification results are on ENVI 5.0 platform post-processing to remove salt and pepper effect to get the final results of desertification hierarchy, as shown in
Taking 1989, 2000, 2009, and 2014 as the time nodes, the research period was divided into three stages, and the area information of desertified land in different time nodes and periods was counted, respectively. The results were shown in
Year | Mild |
Moderate |
Severe |
Serious |
Total | |
---|---|---|---|---|---|---|
1989 | area/km2 | 341.84 | 66.34 | 33.32 | 17.23 | 458.73 |
proportion/% | 74.52 | 14.46 | 7.26 | 3.76 | 100.00 | |
2000 | area/km2 | 1040.19 | 398.26 | 51.51 | 69.98 | 1559.95 |
proportion/% | 66.68 | 25.53 | 3.30 | 4.49 | 100.00 | |
2009 | area/km2 | 983.17 | 375.14 | 28.15 | 27.65 | 1414.11 |
proportion/% | 69.53 | 26.53 | 1.99 | 1.96 | 100.00 | |
2014 | area/km2 | 1020.41 | 254.25 | 51.07 | 44.49 | 1370.22 |
proportion/% | 74.47 | 18.56 | 3.73 | 3.25 | 100.00 | |
Area of change /km2 | 1989–2000 | 698.36 | 331.92 | 18.19 | 52.76 | 1101.22 |
2000–2009 | −57.02 | −23.12 | −23.36 | −42.33 | −145.84 | |
2009–2014 | 37.24 | −120.89 | 22.91 | 16.84 | −43.89 |
From 2000 to 2009, the total area of desertified land decreased by 145.84 km2. This may be related to the formulation of regional desertification control policies, the active implementation of desertification control measures (sand fixation and grass planting, sand sealing and grassland breeding, returning farmland to grassland, rational grazing, etc.), and the enhancement of human environmental awareness. During this period, the area of different degrees of desertified land decreased, and the area of mild desertified land decreased the most, reaching 57.02 km2. These changes show that the measures to control desertification in this period have achieved initial results.
From 2009 to 2014, the total area of desertified land in the study area was still decreasing, but the decreasing rate was significantly slower, and the total area of desertified land decreased only 43.89 km2 during this period. However, according to the area change of different degree desertified land types, the area of moderate desertified land decreased significantly during this period, but the area of other degree desertified land increased. However, the overall trend is gradually improving, indicating that the ecological environment quality of the study area has been significantly improved.
According to the transfer matrix of desertified land of different grades in three periods (
Year | Land type | Mild |
Moderate |
Severe |
Serious |
No |
Total |
---|---|---|---|---|---|---|---|
1989~2000 | Mild |
114.36 | 165.52 | 20.32 | 5.80 | 31.79 | 337.79 |
Moderate |
11.99 | 13.83 | 10.40 | 11.06 | 17.26 | 64.54 | |
Severe |
5.38 | 3.25 | 2.75 | 8.39 | 12.73 | 32.49 | |
Serious |
1.73 | 2.97 | 2.06 | 9.03 | 0.67 | 16.48 | |
No |
903.77 | 210.14 | 14.73 | 29.88 | 1032.36 | 2190.88 | |
Total | 1037.23 | 395.72 | 50.26 | 64.15 | 1094.82 | 2642.19 | |
2000~2009 | Mild |
693.30 | 112.57 | 1.90 | 2.44 | 229.47 | 1039.68 |
Moderate |
139.06 | 226.21 | 11.45 | 4.37 | 16.47 | 397.56 | |
Severe |
13.72 | 17.85 | 9.56 | 3.80 | 6.11 | 51.03 | |
Serious |
31.87 | 10.86 | 4.71 | 15.24 | 3.92 | 66.59 | |
No |
97.86 | 6.14 | 0.26 | 0.37 | 991.46 | 1096.08 | |
Total | 975.81 | 373.62 | 27.88 | 26.22 | 1247.41 | 2650.94 | |
2009~2014 | Mild |
685.30 | 57.69 | 9.20 | 4.87 | 212.98 | 970.02 |
Moderate |
206.29 | 143.79 | 12.19 | 6.73 | 1.59 | 370.58 | |
Severe |
1.28 | 12.21 | 7.78 | 6.01 | 0.05 | 27.32 | |
Serious |
0.29 | 1.73 | 2.50 | 20.44 | 0.02 | 24.98 | |
No |
127.26 | 38.80 | 19.34 | 1.94 | 1023.63 | 1210.96 | |
Total | 1020.41 | 254.21 | 51.00 | 39.99 | 1238.26 | 2603.87 |
From 2000 to 2009, the total area of non-desertified land converted to desertified land was 104.62 km2, and the total area of desertified land converted to non-desertified land was 255.95 km2, indicating that the trend of land change in this period was reverse desertification and the ecological status was getting better. The area of mild desertification land converted to moderate desertification land was 112.57 km2, which was much larger than that of severe desertification land and severe desertification land. However, the area of moderate desertification land was 139.06 km2, and the area of moderate desertification land was 139.06 km2, and the area of moderate desertification land was 139.06 km2, and the area of moderate desertification land was 139.06 km2. The area of severe desertification land was 13.72 km2 and 17.85 km2, respectively. The severe desertification land was mainly converted from moderate desertification land, accounting for 62.5% of the total converted to severe desertification land. There was a distinct improvement in serious desertification land, into the largest proportion, mild desertified land conversion area of 31.87 km2, followed by moderate desertification, 10.86 km2, into the proportion of the severe desertification and desertification land types and smaller, similar serious desertification development is mainly by moderate desertification land conversion. It was 4.37 km2, followed by severe desertification land and mild desertification land, and finally by non-desertification land, which was only 0.37 km2.
From 2009 to 2014, the area of moderately desertified land decreased by 120.89 km2, which was mainly reversed to mild desertified land (206.29 km2). Among the moderate desertified land types, mild desertified land accounted for 52%, followed by non-desertified land (35%). The area of mild desertified land was mainly reversed to non-desertified land, the area of mild desertified land was 57.69 km2, the area of severely desertified land was 9.2 km2 and the area of severely desertified land was 4.87 km2. Severely desertified land was mainly converted from non-desertified land (19.34 km2), and moderate desertified land (12.19 km2), followed by severe desertified land (6.01 km2), developed land (19.34 km2) and moderate desertified land (12.19 km2). Severely desertification land was mainly converted from moderate desertification land and severe desertification land, and the area of serious desertification land was 6.73 km2 and 6.01 km2 respectively. The non-desertified land was mainly converted to mild desertified land (127.26 km2), followed by moderate desertified land (38.80 km2), and the increase of non-desertified land was mainly caused by the reversal of mild desertified land (212.98 km2).In general, desertification showed a slight reversal trend, and the variation trend was slightly different among different degrees of desertification, and desertification intensified in some areas.
Based on the classification results of remote sensing images, desertification information in the study area from 1988 to 2014 was extracted and statistically analyzed. For incomplete data in some years, linear interpolation method was used to complete the data, as shown in
Year | Mild |
Moderate |
Severe |
Serious |
Total area of |
---|---|---|---|---|---|
1989 | 341.84 | 66.34 | 33.32 | 17.23 | 458.73 |
1990 | 1550.73 | 619.49 | 36.83 | 23.06 | 2230.10 |
1991 | 988.16 | 463.25 | 52.88 | 35.69 | 1539.97 |
1992 | 989.64 | 470.52 | 55.75 | 36.30 | 1552.21 |
1993 | 1037.24 | 178.07 | 42.14 | 15.71 | 1273.16 |
1994 | 992.61 | 485.06 | 61.49 | 37.53 | 1576.69 |
1995 | 902.70 | 666.77 | 84.27 | 92.30 | 1746.05 |
1996 | 1259.86 | 915.32 | 56.94 | 22.80 | 2254.91 |
1997 | 1129.57 | 574.31 | 76.70 | 26.73 | 1807.32 |
1998 | 1099.74 | 711.29 | 51.46 | 21.46 | 1883.94 |
1999 | 855.19 | 101.18 | 53.62 | 15.19 | 1025.17 |
2000 | 1040.19 | 398.26 | 51.51 | 69.98 | 1559.95 |
2001 | 758.65 | 885.66 | 299.09 | 152.81 | 2096.21 |
2002 | 1032.52 | 455.95 | 69.41 | 37.52 | 1595.39 |
2003 | 1004.07 | 642.04 | 59.01 | 35.94 | 1741.06 |
2004 | 1138.63 | 715.58 | 63.97 | 31.39 | 1949.56 |
2005 | 1107.66 | 353.78 | 28.01 | 29.88 | 1519.33 |
2006 | 756.22 | 743.77 | 82.84 | 37.78 | 1620.62 |
2007 | 1011.91 | 579.55 | 98.82 | 45.48 | 1735.77 |
2008 | 827.47 | 823.96 | 246.48 | 62.13 | 1960.03 |
2009 | 983.17 | 375.14 | 28.15 | 27.65 | 1414.11 |
2010 | 1046.54 | 259.99 | 62.29 | 24.11 | 1392.93 |
2011 | 1017.85 | 608.63 | 110.31 | 47.93 | 1784.71 |
2012 | 1019.33 | 615.90 | 113.18 | 48.54 | 1796.95 |
2013 | 1185.45 | 1064.87 | 189.23 | 55.39 | 2494.95 |
2014 | 1020.41 | 254.25 | 51.07 | 44.49 | 1370.22 |
As can be seen from the variation trend of the total area of desertified land in the study area (
As can be seen from the area changes of different grades of desertified land (
The increment of the desertified land area can well reflect the change amplitude and trend of desertified land area. The increment information of different types of desertified land area in the study area from 1990 to 2014 is shown in
Year | Mild |
Moderate |
Severe |
Serious |
Total area of |
---|---|---|---|---|---|
1990 | 1208.89 | 553.15 | 3.50 | 5.84 | 1771.38 |
1991 | −562.57 | −156.24 | 16.05 | 12.63 | −690.13 |
1992 | 1.48 | 7.27 | 2.87 | 0.61 | 12.24 |
1993 | 47.60 | −292.45 | −13.61 | −20.59 | −279.05 |
1994 | −44.63 | 306.98 | 19.35 | 21.82 | 303.52 |
1995 | −89.91 | 181.72 | 22.78 | 54.78 | 169.37 |
1996 | 357.16 | 248.54 | −27.33 | −69.51 | 508.86 |
1997 | −130.30 | −341.00 | 19.76 | 3.94 | −447.60 |
1998 | −29.83 | 136.97 | −25.25 | −5.28 | 76.62 |
1999 | −244.55 | −610.11 | 2.16 | −6.27 | −858.77 |
2000 | 185.00 | 297.08 | −2.10 | 54.80 | 534.78 |
2001 | −281.55 | 487.40 | 247.57 | 82.83 | 536.26 |
2002 | 273.87 | −429.71 | −229.68 | −115.29 | −500.81 |
2003 | −28.45 | 186.09 | −10.40 | −1.58 | 145.67 |
2004 | 134.55 | 73.54 | 4.96 | −4.55 | 208.50 |
2005 | −30.97 | −361.80 | −35.96 | −1.51 | −430.24 |
2006 | −351.44 | 389.99 | 54.84 | 7.90 | 101.29 |
2007 | 255.69 | −164.22 | 15.98 | 7.70 | 115.15 |
2008 | −184.44 | 244.41 | 147.65 | 16.65 | 224.27 |
2009 | 155.71 | −448.82 | −218.33 | −34.48 | −545.92 |
2010 | 63.36 | −115.14 | 34.14 | −3.55 | −21.18 |
2011 | −28.69 | 348.63 | 48.02 | 23.82 | 391.79 |
2012 | 1.48 | 7.27 | 2.87 | 0.61 | 12.24 |
2013 | 166.12 | 448.98 | 76.05 | 6.85 | 698.00 |
2014 | −165.04 | −810.63 | −138.16 | −10.90 | −1124.73 |
In general, the increment of the total area of desertification in the study area showed a decreasing trend year by year (
As shown in
In this study, the lake basin area of Gangcha County as the study area was used to analysis the trend of desertification. The multi-phase remote sensing image of the study area was interpreted by RS technology, and the desertification land information of the study area in the past 26 years was extracted. The Desertified land in the area is divided into four grades: light desertification, moderate desertification, heavy desertification, and severe desertification by decision tree classification and statistics on the area and spatial distribution of different degrees of desertification land in the study area. The results of desertification data analysis show that since 1989, the overall desertification in the study area has experienced a fluctuating process of first deterioration and then improvement. The 1990s was the stage of rapid development of desertification in the study area. During this period, the total area of desertification increased by 1101.22 km2. The land area has shown a significant increase trend, of which the lightly desertified land has the largest increase, and the expansion area has reached 698.36 km2. Since the 21st century, the overall desertification in the study area has shown a gradual improvement trend, but local desertification is still aggravating. Through the analysis of the data of the desertified land in the study area over the years, it is found that the total area of desertification in the study area showed a significant increase from 1989 to 2014, with an increased rate of 12.24 km2/y. The area of desertification land in different degrees has an increasing trend, but the increase rate is different. Among them, the area of moderately desertified land has the largest spread rate, which is +7.27 km2/y; followed by heavily desertified land and lightly desertified land, with rates of +2.87 km2/y and 1.48 km2/y, the change rate of the severely desertified land area is the smallest, which is +0.61 km2/y. The increase in the area of different types of desertification showed a downward trend, indicating that the increase in the area of desertification in the study area was slowing down. In 1990 and 2002, the increase in desertification was generally large, and there was an obvious mutation effect.
This work was supported by the National Natural Science Foundation of China “Study on the dynamic mechanism of grassland ecosystem response to climate change in Qinghai Plateau” under grant number U20A2098.