Spatial-temporal evolution characteristics of land use and its driving factor analysis during the period of 2000-2020 in Putian City, Fujian Province

Exploring the spatial and temporal evolution characteristics of land use and its driving force can provide scientific basis for the construction of regional ecological civilization. Based on the remote sensing cloud computing platform of Google Earth Engine (GEE), the study took the remote sensing images of the three periods of 2000, 2010 and 2020 as the basic data and interpreted them to obtain the basic land use data of Putian city and its five districts and counties. The land-use change characteristics of Putian city in those periods were systematically analyzed by using the methods of single land-use dynamics and Geo-informatic Tupu, so as to reveal the Tupu characteristics of land-use transfer and the spatial-temporal evolution pattern in the past 20 years. Then social and economic development indicators were selected to further explore the key driving force of land-use evolution in the study area by canonical correspondence analysis. The results showed that: 1) the land use structure of Putian city was mainly composed of cultivated land and forest land, and the other land types were embedded in them and the built-up land continued to expand outward. 2) The spatial distribution of Tupu units of land use transfer in Putian city was significantly different, change types of land use were diversified and the areas of cultivated land and grassland transferred out were the most obvious. Different degrees of cultivated land abandonment had occurred in various regions. Putian city was facing great challenges in ensuring food security and curbing the non-agricultural and non grain of cultivated land. 3) The macro-economic development in a specific period, especially the urban expansion and the development of the secondary industry caused by the merger of cities and counties, were the key factors driving the spatial and temporal evolution of land use and the differential distribution pattern in Putian city. The author suggested that land resources should be used efficiently in addition to optimizing the layout of land use and carrying out the renovation of abandoned land. Therefore, the study proposed to strengthen the scientific planning and effective regulation of land use, combine spatial and temporal resources of cultivated land with agriculture, culture and tourism, and build ecological barriers and ecological corridors to ensure the coordinated and sustainable development of regional ecology and economy.


Introduction
Land use / cover change (LUCC) is the most direct manifestation in the impact of human 32 activities on the earth's surface system, and has become one of important research areas in global 33 environmental change and sustainable development. With the advancement of the rapid 34 globalization, urbanization and industrialization in the world, especially in developing countries, the 35 social and economic transformation has also had a significant impact on regional land use [1,2], 36 which makes land use transformation become a new theme and frontier of LUCC research. 37 Since the 1990s, China's urban-rural transformation has led to a drastic land-use transition 38 (LUT), which has greatly affected the ecosystem structure and its service functions, threatening the 39 sustainable development of human beings and the sustainable supply of ecosystem services [3][4][5]. 40 For example, the transformations of native grasslands, ecological forests and wetlands into 41 farmland, artificial forests and built-up land, have greatly promoted a substantial increase in the variables is called feature random subspace; 4) integrate the prediction results of N decision trees 147 and determine new sample categories by voting; 5) through the above steps, an RF model with N 148 decision trees is established, its performance is scored by the test set, and the importance of each 149 feature is ranked [31]. (2) 159 EVI is considered as an improved NDVI, which increases the sensitivity and vegetation 160 monitoring ability in high biomass areas by decoupling the canopy background signal and reducing 161 atmospheric impact. EVI was adopted by MODIS (modern resolution imaging spectroradiometer) 162 terrestrial discipline group as the second global vegetation index to monitor terrestrial 163 photosynthetic vegetation activities. The calculation formula is as follows [34]: 164 (3) 165 Where ρ NIR 、ρ RED 、ρ SWIR and ρ BLUE are the reflectance of near-infrared, red, short wave infrared 166 and blue in Landsat images.

167
Single land-use dynamic index 168 Single land-use dynamic index can quantitatively describe the speed of regional land use 169 change, and play an important role in comparing the regional differences of land use transformation 170 and analyzing the trend of land use change [4] . The formula is as follows: 171 (4) 172 Where K refers to single land-use dynamic index, U a and U b represent the area of a certain land 173 type at the beginning and the end of the study respectively, and T denotes the research period.

175
Geo-informatic Tupu is a spatio-temporal analysis method that records land use change with 176 Tupu units [20]. It has the dual properties of graph and pedigree. The graph and the pedigree are respectively used to represent spatial location characteristics and process variation [22] The study selected seven types of land use in Putian city and its five districts and counties as 215 species variables. According to the regional socio-economic development trend and data 216 availability, 17 socio-economic indicators with extremely significant correlation with land use 217 change were selected as the environmental variables, namely, fishery output value, the gross value 218 of primary industry, GDP, total industrial output value, urbanization rate, the gross value of tertiary 219 industry, the gross value of secondary industry, the area of economic crops, the non-agricultural population, the total retail sales of social consumer goods, the total population, and the agricultural 221 population, which basically include the main socioeconomic indicators of the region. 222 CANOCO 4.5 and CANOD RAW 4.5 were used for CCA analysis. 223

224
Analysis of land use structure change 225 The status map of land use classification in Putian City in 2000, 2010 and 2020 was obtained 226 based on the GEE platform (Fig 1). According to formula (7) and formula (8) forest land area at the three-time courses exceeds more than 50% of the total area of the study area;

252
The land use dynamic index can be also used to estimate the change of land use structure in a 253 specific period from another aspect. The results in Table 1  From the distribution results of land use (Fig 2), the study found that the types of land use  Table 2). The spatial distribution of these three types of Tupu units in the study area 285 was a large-scale high-density distribution (Fig 3a). The Tupu units of "cultivated land→forest land 286 "(code 12) and" forest land → cultivated land" (code 21) accounted for 6.70% and 6.03% of all 287 converted land use types, respectively. The Tupu units of "cultivated land→water body" (code 15) 288 and "cultivated land→grassland" (code 14) were also noteworthy, accounting for 4.11% and 3.34% 289 of the total converted land use types, respectively. These land use types were widely distributed in 290 the south and coastal areas of the study area. 291

295
Further from the transformation area of each land type ( km 2 ), respectively. The transferred area of forest land was 14.10 km 2 , mainly from cultivated land 315 and orchard land. The transferred-out area of grassland (9.01 km 2 ) was smaller than the transferred 316 in area (7.73 km 2 ), with a difference of 1.28 km 2 . The grass land transferred out mainly converted to cultivated land (5.35 km 2 ), and the area transferred in mainly came from cultivated land (7.49 km 2 ). 318 The conversion of cultivated land to grassland was mainly due to abandoned cultivated land, which 319 finally evolved into grassland, and most of the grassland was finally transferred into built-up land. 320 321 The difference of land use structure transformation in Putian city was studied by generating the 327 rising Tupu of Putian city in 2000-2010 and 2010-2020 (Fig 4) and making statistics and analysis 328 on the transferred data (Table 4). 329 As shown in Table 4   The temporal and spatial evolution of land use change in Putian city was studied 376 by generating the falling Tupu of Putian city in 2000-2010 and 2010-2020 (Fig 5) and 377 making statistics and analysis on the transferred data (Table 5). 378 According to Table 5 (Table 7), there were 6 indicators with extremely significant differences. The larger r 2 440 and smaller Pr indicated the index was more important. The order of importance was 441 the gross value of the secondary industry, the output value of forestry, the total industrial output value, the gross domestic product, the total retail sales of social 443 consumer goods, and the gross value of the tertiary industry. In CCA ranking (Fig 6),