Research on changes and prediction of coastal urban landscape pattern based on RF-CA-Markov model: Taking Lianyungang City as an example

doi:10.14088/j.cnki.issn0439-8114.2022.13.020

Abstract

Abstract: Lianyungang City was selected as the study area，and the random forest algorithm was used to measure the driving effects between various factors and landscape types. A cellular automata model combining the Markov method was built to simulate the land⁃scape pattern change in coastal cities，and the future landscape pattern of the study area was simulated and predicted. The results showed that the random forest model had a strong ability to explain the distribution of all landscape types in Lianyungang City，and the ROC value of all landscape types was greater than 0.800，among which the ROC value of cultivated land，forest land and construction land were more than 0.900. The distribution suitability of landscape types in Lianyungang City was influenced by natural and human factors. The RF-CA-Markov model could well simulate the landscape pattern characteristics of Lianyungang City，and the Kappa in⁃dex was 0.946 1. From 2010 to 2025，the area of construction land in Lianyungang continues to increase，mainly concentrated in the marginal areas of cities and towns. The areas of cultivated land，salt pan and forest land continues to decrease. The abandoned salt pans were mainly transferred to the artificial pond，and some of them were turned into unused land. The degree of fragmentation of landscape pattern increased，the connectivity between the same landscape decreased，the landscape diversity and uniformity in⁃creased，and the overall shape tended to be complex.

Key words: landscape pattern, time and space changes, RF-CA-Markov model, Lianyungang City

CLC Number:

P901
F301.24

SUN Jing-ya, WANG Xu, CHEN Kun-lun. Research on changes and prediction of coastal urban landscape pattern based on RF-CA-Markov model: Taking Lianyungang City as an example[J]. HUBEI AGRICULTURAL SCIENCES, 2022, 61(13): 106-112.

References

[1] MANDELBROT B.How long is the coast of Britain?Statistical self-similarity and fractional dimension[J].Science,1967,156(3775):636-638.
[2] 马万栋,张渊智,施平,等.海岸带土地利用/土地覆被变化研究进展[J].地理科学进展,2008(5):87-94.
[3] 王倩. 防城港海岸带景观格局变化研究[D].南宁:广西大学,2013.
[4] 索安宁,关道明,孙永光,等.景观生态学在海岸带地区的研究进展[J].生态学报,2016, 36(11): 3167-3175.
[5] 巫丽芸,何东进,游巍斌,等.东山岛海岸带景观破碎化时空梯度分析[J].生态学报,2020,40(3): 1055-1064.
[6] 王奇,李伟芳,寇相玮,等.甬台温海岸带景观格局及生态脆弱性时空演变特征[J].水土保持通报,2020,40(1): 297-303,325.
[7] COSTANZA R, VOINOV A.Landscape simulation modeling: A spatially explicit, dynamic approach[M]. Dordrecht: Springer science & business media, 2003.
[8] 刘世梁,安南南,尹艺洁,等.广西滨海区域景观格局分析及土地利用变化预测[J].生态学报,2017,37(18): 5915-5923.
[9] 潘润秋,罗启源,肖迪,等.基于CLUE-S模型的深圳海岸带土地利用变化模拟[J].测绘与空间地理信息,2016,39(4):32-36,40.
[10] 刘菁华,李伟峰,周伟奇,等.京津冀城市群景观格局变化机制与预测[J].生态学报, 2017, 37(16): 5324-5333.
[11] 何丹,金凤君,周璟.基于Logistic-CA-Markov的土地利用景观格局变化——以京津冀都市圈为例[J].地理科学,2011,31(8): 903-910.
[12] 张学儒,周杰,李梦梅.基于土地利用格局重建的区域生境质量时空变化分析[J].地理学报,2020,75(1): 160-178.
[13] BREIMAN L.Random forests[J]. Machine learning,2001,45(1):5-32.
[14] KNOL M J, LE CESSIE S, ALGRA A, et al.Overestimation of risk ratios by odds ratios in trials and cohort studies: Alternatives to logistic regression[J]. Canadian medical association journal, 2012, 184(8): 895-899.
[15] 李欣海. 随机森林模型在分类与回归分析中的应用[J].应用昆虫学报, 2013, 50(4): 1190-1197.
[16] 邬建国. 《景观生态学》:格局、过程、尺度与等级[M].北京:高等教育出版社,2007.
[17] 胡碧松,张涵玥.基于CA-Markov模型的鄱阳湖区土地利用变化模拟研究[J].长江流域资源与环境,2018, 27(6): 1207-1219.
[18] SANG L, ZHANG C, YANG J, et al.Simulation of land use spatial pattern of towns and villages based on CA-Markov model[J]. Mathematical and computer modelling, 2011, 54(3-4):938-943.
[19] CUTLER D R, EDWARDS T C, BEARD K H, et al.Random forests for classification in ecology[J]. Ecology,2007,88(11): 2783-2792.
[20] FAWCETT T.An introduction to ROC analysis[J]. Pattern recognition letters, 2006, 27(8):861-874.
[21] 罗紫薇,胡希军,韦宝婧,等.基于多准则CA-Markov模型的城市景观格局演变与预测——以上杭县城区为例[J].经济地理, 2020, 40(10): 58-66.