Efficiency of Google Earth Engine (GEE) system in land use change assessment and predicting it using CA-Markov model (Case study of Urmia plain)

Subject Areas : Spatial data infrastructures and standardisation

Naser Soltani ¹ ( Assistant Professor, Department of Geography, Faculty of Literature and Humanities, University of Urmia, Urmia, Iran )
Vahid Mohammadnejad ² ( Assistant Professor, Department of Geography, Faculty of Literature and Humanities, University of Urmia, Urmia, Iran )

Received: 2020-12-23 Accepted : 2021-01-22 Published : 2021-09-23

Keywords: Google Earth Engine, Classification, Landuse, Urmia,

Abstract :

Background and ObjectiveLand use reflects the interactive features between humans and the environment and describes how humans are exploited for one or more purposes on earth. Land use is usually defined based on human use of land, with an emphasis on the functional role of land in economic activities. Land use map is one of the main factors in the study of natural resources and environmental management. Knowing the changes in land use and examining their causes and factors in a period of time can be of interest to planners and managers. The use of satellite data is a good tool for land use mapping, especially in large geographical areas, due to the provision of a wide and integrated view of an area, reproducibility, easy access, high accuracy of information obtained, and high-speed analysis. One of the most widely used methods of extracting information from satellite images is image classification, which allows users to generate different information. Google Earth Engine (GEE) is a web, cloud-based system developed by Google to store and analyze large amounts of data at the petabyte scale (including various satellite imagery, digital models, climatic and vector data). Speed in processing and access to diverse data is one of the issues and problems of land use change studies. The purpose of this paper is to classify satellite images using the support vector machine learning method in the two periods of 2000 and 2020 and to produce a land use map of these two periods in the Google Earth engine system. Materials and Methods In this paper, Urmia city and its surrounding areas (Urmia plain) have been evaluated. In order to prepare land use maps and study its changes, Landsat 7 ETM+ sensor for 2000 and Landsat OLI 8 for 2020 have been used. Images from June were used, when vegetation reached its maximum vegetative growth. Various methods have been developed to monitor and measure land cover and land use changes. In this paper, the efficiency of the Google Earth Engine system for collecting, managing, and processing remote sensing data has been evaluated in order to prove and introduce the speed and accuracy of this system. In order to produce the land use map, the Support Vector Machine classification method has been used. The main difference between this paper and other research is that the management and processing of images have been done in the Google Earth Engine system, which means that the researcher does not need expensive and licensed software such as ENVI and only by access to the Internet can do the processing. By developing the code for image classification using the support vector machine method, the images of 2000 and 2020 were classified. Six land use classes were identified, including barren lands, man-made lands, orchards, irrigated agriculture, rainfed agriculture, and irrigated areas. After classifying images, the results were stored in Google Drive and prepared for further analysis. The classification results were entered into ArcGIS software and the classification accuracy was evaluated using control points obtained from Google Earth images as well as data related to the land use management plan of West Azerbaijan province. In this paper, in addition to preparing a land use map in the Google Earth Engine system, it was used to forecast and model land uses for 2040 using the CA-Markov transfer estimator. Results and Discussion After calling and classification of images in the Google Earth engine environment using the SVM method, land use map for 2000 and 2020 was produced. The prepared maps include man-made lands, orchards, irrigated agriculture, rainfed agriculture, and barren lands. A comparison of different land use in 2000 and 2020 shows that extensive changes have taken place in them. Some of these changes are positive and some are negative. The area of barren lands in 2020 compared to 2000 has increased by about 10 square kilometers, man-made lands, 42.62 square kilometers, orchards 67 square kilometers, and water bodies 0.39 square kilometers. In contrast, rainfed agriculture has lost 39.45 and irrigated agriculture has lost 80 square kilometers. The reason for the increase in orchards can be seen in the change of irrigated agricultural uses to orchards, as well as urban development and the creation of various human infrastructures, which is very evident in recent years. Most of the changes are related to the use of orchards with a positive trend during which many irrigated agricultural lands have become garden lands. These changes have increased the production of horticultural products in Urmia and become one of the hubs of horticultural production, especially apples. the area of man-made land has almost doubled, which usually happens in other parts of the country and is normal. Usually, with the increase in the population of cities as well as villages and the need to build new buildings and infrastructure facilities such as factories, sports fields, roads, entertainment spaces, etc., man-made uses have increased. According to the forecast for 2040 using the CA-Markov method in Idrisi software, the highest growth is related to rainfed agricultural use. It is predicted that during this period, the area of rainfed lands will reach 73.40 square kilometers. The man-made land will increase to 90.9 square kilometers. While its value in 2020 was 76.38 square kilometers. On the other hand, the area of orchards will increase from 31.61 square kilometers in 2020 to 72.15 square kilometers. Irrigated agriculture will increase to 27.38 square kilometers with an increasing trend. Conclusion Studies show that the growth of man-made lands in Urmia city and its surroundings is not commensurate with other land uses and this has led to the growth of land use area of ​​the man-made lands compared to other uses and this issue has caused the phenomenon of expansion has become in Urmia city. On the other hand, the results show that the study of land use using the time series of satellite images is a time saver and cost, and as mentioned in the paper. different land uses for the years 2000 and 2020, prepared using the Google Earth system, and their changes were identified. Another important result of this paper is the high efficiency of the GEE system in processing large volumes of satellite images. Using this system does not require any specialized remote sensing software and the user can easily process various data using a computer browser or even a smartphone. Another important point is that in this system, there is no need to download different images, but the user can only download the processing result. This is very useful in terms of time and processing speed. The GEE system is able to process large volumes of time series data (here satellite imagery), different regions of the world with very high speed and very low time, and present the results in the form of various maps and graphs.

References:

Afify HA. 2011. Evaluation of change detection techniques for monitoring land-cover changes: A case study in new Burg El-Arab area. Alexandria Engineering Journal, 50(2): 187-195. doi:https://doi.org/10.1080/014311698216062.

Asghari S, Mohammadnejad V, Emami H. 2019. Investigation land use change with use of a pixel-based method and object-oriented method and analysis of the effect of land use change on soil erosion (Case study of Maragheh county) Quantitative Geomorphological Researches, 29(8): 160-178. (In Persian).

Burges CJ. 1998. A tutorial on support vector machines for pattern recognition. Data Mining and Knowledge Discovery, 2(2): 121-167. doi:https://doi.org/10.1023/A:1009715923555.

Civco DL, Hurd JD, Wilson EH, Arnold CL, Prisloe Jr MP. 2002. Quantifying and describing urbanizing landscapes in the Northeast United States. Photogrammetric Engineering and Remote Sensing, 68(10): 1083-1090. http://worldcat.org/issn/00991112.

Deng C, Wu C. 2012. BCI: A biophysical composition index for remote sensing of urban environments. Remote Sensing of Environment, 127: 247-259. doi:https://doi.org/10.1016/j.rse.2012.09.009.

Goldblatt R, You W, Hanson G, Khandelwal AK. 2016. Detecting the boundaries of urban areas in india: A dataset for pixel-based image classification in google earth engine. Remote Sensing, 8(8): 634. doi:https://doi.org/10.3390/rs8080634.

Hansen MC, Loveland TR. 2012. A review of large area monitoring of land cover change using Landsat data. Remote sensing of Environment, 122: 66-74. doi:https://doi.org/10.1016/j.rse.2011.08.024.

Hartter J, Lucas C, Gaughan AE, Aranda LL. 2008. Detecting tropical dry forest succession in a shifting cultivation mosaic of the Yucatán Peninsula, Mexico. Applied Geography, 28(2): 134-149. doi:https://doi.org/10.1016/j.apgeog.2007.07.013.

Herold M, Scepan J, Clarke KC. 2002. The use of remote sensing and landscape metrics to describe structures and changes in urban land uses. Environment and planning A, 34(8): 1443-1458. doi:https://doi.org/10.1068/a3496.

Heydarian P, Rangzan K, Maleki S, Taghizadeh A. 2014. Land use change detection using post classification comparison LandSat satellite images (Case study: land of Tehran). Journal of RS and GIS for Natural Resources, 4(4): 1-10. http://girs.iaubushehr.ac.ir/article_516552_en.html. (In Persian).

Johnson RD, Kasischke E. 1998. Change vector analysis: A technique for the multispectral monitoring of land cover and condition. International Journal of Remote Sensing, 19(3): 411-426. doi:https://doi.org/10.1080/014311698216062.

Keshavarz A, Ghasemiyan H. 2005. A fast algorithm based on support vector machine for classification of hyperspectral images using spatial correlation. Iranian Journal of Electrical Engineering and Computer Engineering, 3: 44-37. (In Persian).

Li G, Lu D, Moran E, Hetrick S. 2013. Mapping impervious surface area in the Brazilian Amazon using Landsat Imagery. GIScience & Remote Sensing, 50(2): 172-183. doi:https://doi.org/10.1080/15481603.2013.780452.

Liu C, Shao Z, Chen M, Luo H. 2013. MNDISI: a multi-source composition index for impervious surface area estimation at the individual city scale. Remote Sensing Letters, 4(8): 803-812. doi:https://doi.org/10.1080/2150704X.2013.798710.

Lu D, Mausel P, Brondizio E, Moran E. 2004. Change detection techniques. International Journal of Remote Sensing, 25(12): 2365-2401. doi:https://doi.org/10.1080/0143116031000139863.

Lu D, Weng Q. 2004. Spectral mixture analysis of the urban landscape in Indianapolis with Landsat ETM+ imagery. Photogrammetric Engineering & Remote Sensing, 70(9): 1053-1062. doi:https://doi.org/10.14358/PERS.70.9.1053.

Mokhtari M, Najafi A. 2015. Comparison of support vector machine and neural network classification methods in land use information extraction through Landsat TM data. Journal of Science and Technology of Agriculture and Natural Resources, 19(72): 35-45. doi:https://doi.org/10.1080/014311698216062. (In Persian).

Mountrakis G, Im J, Ogole C. 2011. Support vector machines in remote sensing: A review. ISPRS Journal of Photogrammetry and Remote Sensing, 66(3): 247-259. doi:https://doi.org/10.1016/j.isprsjprs.2010.11.001.

Patel NN, Angiuli E, Gamba P, Gaughan A, Lisini G, Stevens FR, Tatem AJ, Trianni G. 2015. Multitemporal settlement and population mapping from Landsat using Google Earth Engine. International Journal of Applied Earth Observation and Geoinformation, 35: 199-208. doi:https://doi.org/10.1016/j.jag.2014.09.005.

Pourahmad A, Seifodini F, Parnon Z. 2011. Migration and land use change in Islamshahr city. Arid Regions Geographic Studies, 2(5): 131-150. https://www.sid.ir/en/journal/ViewPaper.aspx?id=251656. (In Persian).

Schneider A, Friedl MA, Potere D. 2010. Mapping global urban areas using MODIS 500-m data: New methods and datasets based on ‘urban ecoregions’. Remote Sensing of Environment, 114(8): 1733-1746. doi:https://doi.org/10.1016/j.rse.2010.03.003.

Seto KC, Fragkias M, Güneralp B, Reilly MK. 2011. A meta-analysis of global urban land expansion. PloS one, 6(8): e23777. doi:https://doi.org/10.1371/journal.pone.0023777.

Shao Y, Li GL, Guenther E, Campbell JB. 2015. Evaluation of topographic correction on subpixel impervious cover mapping with CBERS-2B data. IEEE Geoscience and Remote Sensing Letters, 12(8): 1675-1679. doi:https://doi.org/10.1109/LGRS.2015.2419135.

Shelestov A, Lavreniuk M, Kussul N, Novikov A, Skakun S. 2017. Exploring Google Earth Engine platform for big data processing: Classification of multi-temporal satellite imagery for crop mapping. Frontiers in Earth Science, 5(7): 1-17. doi:https://doi.org/10.3389/feart.2017.00017.

Soffianian AR, Khodakarami L. 2011. Land Use Mapping Using Fuzzy Classification: Case Study in Three Catchment Areas in Hamedan Province. Town and Country Planning, 3(4): 95-114. https://jtcp.ut.ac.ir/article_23206_23200.html?lang=en. (In Persian).

Sun Z, Guo H, Li X, Lu L, Du X. 2011. Estimating urban impervious surfaces from Landsat-5 TM imagery using multilayer perceptron neural network and support vector machine. Journal of Applied Remote Sensing, 5(1): 053501. doi:https://doi.org/10.1117/1.3539767.

Sun Z, Xu R, Du W, Wang L, Lu D. 2019. High-resolution urban land mapping in China from sentinel 1A/2 imagery based on Google Earth Engine. Remote Sensing, 11(7): 752. doi:https://doi.org/10.3390/rs11070752.

Wahap N, Shafri HZ. 2020. Utilization of Google Earth Engine (GEE) for land cover monitoring over Klang Valley, Malaysia. In: IOP Conference Series: Earth and Environmental Science, vol 1. IOP Publishing, pp 012003. https://iopscience.iop.org/article/012010.011088/011755-011315/012540/012001/012003/meta.

Wang Z, Gang C, Li X, Chen Y, Li J. 2015. Application of a normalized difference impervious index (NDII) to extract urban impervious surface features based on Landsat TM images. International Journal of Remote Sensing, 36(4): 1055-1069. doi:https://doi.org/10.1080/01431161.2015.1007250.

Weng Q, Hu X. 2008. Medium spatial resolution satellite imagery for estimating and mapping urban impervious surfaces using LSMA and ANN. IEEE Transactions on Geoscience and Remote Sensing, 46(8): 2397-2406. doi:https://doi.org/10.1109/TGRS.2008.917601.

Wu C, Murray AT. 2003. Estimating impervious surface distribution by spectral mixture analysis. Remote sensing of Environment, 84(4): 493-505. doi:https://doi.org/10.1016/S0034-4257(02)00136-0.

Wu M, Zhao X, Sun Z, Guo H. 2019. A hierarchical multiscale super-pixel-based classification method for extracting urban impervious surface using deep residual network from worldview-2 and LiDAR data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(1): 210-222. doi:https://doi.org/10.1109/JSTARS.2018.2886288.

Xu H. 2010. Analysis of impervious surface and its impact on urban heat environment using the normalized difference impervious surface index (NDISI). Photogrammetric Engineering & Remote Sensing, 76(5): 557-565. doi:https://doi.org/10.14358/PERS.76.5.557.

Yousef S, Tazeh M, Mirzaee S, Moradi H, Tavangar S. 2011. Comparison of different classification algorithms in satellite imagery to produce land use maps (Case study: Noor city). Journal of RS and GIS for Natural Resource, 5(3): 67-76. (In Persian).

Zhang C, Sargent I, Pan X, Li H, Gardiner A, Hare J, Atkinson PM. 2018. An object-based convolutional neural network (OCNN) for urban land use classification. Remote Sensing of Environment, 216: 57-70. doi:https://doi.org/10.1016/j.rse.2018.06.034.