Access the full text.
Sign up today, get DeepDyve free for 14 days.
Environmental & Socio-economic Studies DOI: 10.2478/environ-2022-0015 Environ. Socio.-econ. Stud., 2022, 10, 3: 33-42 ________________________________________________________________________________________________ Original article Relationship between densification and NDVI loss. A study using the Google Earth Engine at local scale Juan Pablo Celemin*, Maria Eugenia Arias Institute of Geography, History and Social Sciences, National Scientific and Technical Research Council, Pinto 399, Tandil, Buenos Aires Province, Argentina E–mail address (*corresponding author): firstname.lastname@example.org ORCID iD: Juan Pablo Celemín: https://orcid.org/0000-0002-8917-8061; María Eugenia Arias: https://orcid.org/0000-0002- 1592-2230 ______________________________________________________________________________________________________________________________________________ A B S T R A C T Latin American cities are amongst those with the highest rates of urbanization in the world. This process has involved their territorial expansion as well as the densification of some of its neighborhoods, in mainly central areas. This is the case of the city of Santiago del Estero (Argentina) that increased its population by 33% between 1991 and 2010 with the consequent transformations of the local space. In this context, this study analyzes the evolution of vegetated areas and densification of the central area of the city using satellite data. We analyzed two indices: normalized difference vegetation index (NDVI) and Urban Index (UI) time-series data, for the 1992–2011 year period, using the Google Earth Engine for processing Landsat 5 TM images. We found that the NDVI showed a decreasing trend in the timelapse under consideration, while the UI performance registered the opposite trend. The mean NDVI decreased from 0.161 (1992) to 0.103 (2011) while the UI mean increased from 0.003 to 0.036 in the same timelapse. Further, the NDVI has a strong negative correlation with UI (R -squared = -0.862). The results are consistent with the census information that recorded an important demographic and housing growth for the entire city in this period. KEY WORDS: vegetation cover loss, densification, NDVI, Google Earth Engine, Santiago del Estero, Argentina ARTICLE HISTORY: received 13 May 2022; received in revised form 24 August 2022; accepted 25 August 2022 ______________________________________________________________________________________________________________________________________________ 1. Introduction environmental impact, diseconomies, and a decrease in the quality of life of the local population. Rapid urban sprawl is a major contributory factor Although urban growth is referred to as necessary for environmental change in many parts of the world. for a sustainable economy, uncontrolled urban Latin America and the Caribbean is the most growth can cause several problems, such as loss of urbanized region in the developing world and is open spaces, landscape modification, environmental characterized by accelerated growth. One of the pollution, traffic congestion, pressure on insufficient key developments that shaped Latin American cities infrastructure, and other social and economic is the migration of people from the countryside to problems (NOLÈ ET AL., 2013). Moreover, in this the city, a phenomenon that has generated regional region, urbanization is one of the main anthropogenic imbalances in most countries of the region (VARGAS- factors that has produced the reduction of green BOLAÑOS ET AL., 2020). Hence, in addition to presenting areas and the replacement of pre-existing habitats in high rates of urbanization, Latin American cities also cities of Latin America (BERKOWITZ ET AL., 2003). display new urban peripheries, usually without To address this context, FERRO (2001) argues that planning, resulting in several issues that have urban planners may consider diverse approaches: characterized the cities of this part of the world: new developments planned in the suburbs (a practice that responds to a "traditional" form of growth), while undeveloped land outside the city could be new cities in the area, or the reuse of existing space retained as a natural environment (EMILSSON & through densification projects. As a substitute to SANG, 2017). Although vegetation cover loss is "traditional" urban sprawl, there is a "densification” prevalent, densification does not always indicate a approach through the re-ordering of large areas reduction in greenness. In certain instances, more that are well located, but that have deteriorated, vegetation in urban areas has emerged from or are misused or vacant. In Latin America, the planning and the drive for more sustainable cities, outcome is an amalgam of growth where the middle such as the cases of Taipei (WANG ET AL., 2018) and upper residential areas have practically and Singapore (GAW & RICHARDS, 2021). On the other disappeared due to increasingly smaller high- hand, most case studies in Argentina demonstrate density buildings, or from the edification of new the opposite results (MERLOTTO ET AL., 2012; PAOLINI constructions that replace homes in old residential ET AL., 2016; FERRELLI & BUSTOS., 2015; ARBOIT & neighborhoods. As opposed to, the lower strata that MAGLIONE, 2018). Therefore, in most cases, this do not have the resources to access the formal process usually involves the loss of green spaces housing market, and are located in the most distant with negative environmental impacts such as the peripheries, in informal urbanization lots where alteration of wind and temperature patterns (PAULEIT they build homes "progressively". Consequently, ET AL., 2005; FONTENELLE ET AL., 2015; LEMONSU ET the central areas of the cities have been condensing AL., 2015). with great force, appealing to the tertiary sector, and Land-use changes are the most notable indicator creating at the same time, an area of commercial of the human footprint and are considered to be activity and services, to which is added the the most important factor of biodiversity loss and construction of residential buildings. Hence, while land degradation. The effect of land-use change varies the central areas are densified, the single-family by region and geographic location (ZURQANI ET AL., neighborhoods on the periphery are also growing in 2019). Generally, in developed countries there is a process that does not necessarily occur in a planned effective territorial planning, while, in Latin America, manner. From the urban economic perspective, it is more anarchic, coexisting urban planning MATTOS, (2016), DEL RÍO (2017) and BENSÚS TALAVERA measures with growth areas without any control. (2018) maintain that this densification course is Given that cities in this region will continue to associated with an intense exploitation of the expand and densify, the study of changes in urban urban land in areas of high construction demand, land cover and their impacts on urban vegetation closely linked to the real estate market. is of fundamental interest (FERRELLI ET AL., 2018). Densification is usually more noticeable in specific Tracking urban land cover change, through the neighborhoods/parts of a city as recent studies interpretation of satellite imagery, can be an indicate (DARCHEN & POITRAS, 2018; TILLIE ET AL., extremely valuable tool for urban planners in 2018; TREIJA ET AL., 2018; SKREDE & BERG, 2019; detecting the effects of environmental change ANDERSSON ET AL., 2020; LI ET AL., 2021; EGGIMANN (HUANG ET AL., 2009). At present there are numerous ET AL., 2021; CELLUCI & SIVO, 2021; among others), satellite platforms that record terrestrial information, and is viewed as an answer to the uncontrolled which is disseminated in different repositories. urban sprawl, however, compared to urban This results not only in a wide variety of data, but growth, densification has some environmental also makes it imperative to handle these large merits, but is not without negative environmental volumes of information more efficiently. Due to impacts (NÆSS ET AL., 2020; COPPOLA, 2012). Dense the high volumes of data captured and archived cities are a logical choice for an increasingly urbanized regularly and over a long period, changes in land world, where worries about environmental use can now be measured not only in two-time sustainability and urban growth are significant snapshots, but continuously over many time intervals (UN HABITAT, 2012). Among their many benefits, (JIANYA ET AL., 2008). Satellite imagery is rapidly dense cities, in addition to helping to preserve being utilized to construct diverse maps of land rural land, also reduce greenhouse gas emissions surface features, such as vegetation, snow, water, (LIBERTUN DE DUREN AND COMPEÁN, 2016). Densification and built-up land characteristics, automatically is viewed as an appropriate measure to cope with and semi-automatically (FIROZJAEI ET AL., 2019). More fast urban growth that aims to limit the expansion of specifically, Landsat sensors have been instrumental the constructed area on the outskirts by expanding in observing geographic phenomena, with datasets additional living space within already built areas going back decades (LI JIAN & ROY, 2017). (BERNDTSSON ET AL., 2019) and can be seen as an Time series analysis of changes in land cover opportunity for sustainable urban development, as allows researchers to understand the general trends it fosters resource efficiency and transportation, and dynamics of land-use changes over complete time series rather than a simple increase, or decrease, more constant throughout the year in any climatic between two points in time. Hence, modeling conditions around the world (KUMARI ET AL., 2020). vegetation change rates as well as built-up area UI was developed as a built-up index in the 1990s evolution can be done more accurately using a to generate data and analyze the situation and sprawl time series provided by satellite imagery (TROMBETTI pattern of built-up areas using remote sensing ET AL., 2008). Historically, the large volumes of data (KAWAMURA ET AL., 1996). The UI normalizes data provided by continuous satellite monitoring the NIR and SWIR 2 bands, making use of the inverse represented a technical challenge for interpretation correlation between NIR and SWIR brightness in and analysis. The processing of large time series constructed areas. This procedure is based on the datasets has been made accessible thanks to fact that urban and bare ground areas have low recent technological advances in computing, in reflectivity in the NIR band but relatively high particular the creation of Google Earth Engine (GEE). reflectivity in the SWIR bands (DO ET AL., 2021). In this context, access to historical and current The UI is computed by using the bands 7 and 4 from remote sensing data using Google Earth Engine's Landsat Thematic Mapper (TM) imagery. It is very geospatial technology represents a significant similar to NDBI and is probably the most used improvement in monitoring and evaluating land- built-up index, with the difference being that it use change over time (ZURQANI ET AL., 2019). uses SWIR2 instead of SWIR1. Vegetation greenness and built-up land are the Although the growth and densification of cities two most important factors for studying urban are two processes that coexist in Latin American variations. The NDVI reflects the health and density cities, most of the research focuses on urban sprawl, of the vegetation, while the built-up indices refer with fewer articles analyzing the densification at to constructed areas. The NDVI and built-up indices intra-urban scale. Consequently, the initial purpose have also been evaluated as predictors and factors in of this research is to use The GEE to process the urban land change (GROVER & SINGH, 2015). Due to large free satellite datasets that are available for the variability and spectral closeness of built-up the long-term monitoring of NDVI and UI in a case areas and bare land, using this type of index is study focused on the downtown neighborhood particularly difficult. Hence, several indices for of the city of Santiago del Estero, in a period mapping built-up and other land cover types in comprising the years 1992 and 2011. The reason urban areas have been used in various studies, for the selection of this timelapse is not arbitrary including the Normalized Difference Built-Up Index since it comprises, approximately, the last three (NDBI) (He et al., 2010), Index-based Built-Up Index censuses carried out in the country (1991, 2001, (IBI) (XU, 2008), Urban Index (UI) (KAWAMURA ET 2010) which provided data, albeit partially, on AL., 1996), Normalized Difference Bareness Index demographic growth as well as on the number of (NDBaI) (ZHAO & CHEN, 2005), and Bare soil index houses in the city. The downtown neighborhood (BSI). The spectral performance of built-up land is the central area that has been more patently and other properties related to wavelengths of the affected by the densification process throughout electromagnetic spectrum in terms of absorption, the city in the last decades. To better contextualize or reflection, serve as the foundation for the this land-change scenario, information on population development of these indices (FIROZJAEI ET AL., 2019), and housing units from these three census periods although each one has its own set of potentials was obtained from the National Statistical and and drawbacks. NDBI and UI, for example, have Census Institute (INDEC) web page. Unfortunately, the limitation of mixing to some degree built-up there is scarce disaggregated data available on and bare land areas, but on the other hand, are that web page, with the exception of the 2011 census. easy to implement (SINHA ET AL., 2016). Moreover, recent studies have demonstrated that using built- 2. Study area up indices, such as NDBI and UI- recurring to the SWIR 1/SWIR 2 bands, are more effective at Santiago del Estero is the capital of the detecting built-up areas, as they possess higher homonymous province in northern Argentina. reflectance values, allowing them to be readily With a surface area of 2,116 km and a population distinguished from other land uses. Therefore, UI, of 252,192. It is the country's tenth largest city, like NDBI, is a relevant indicator for urban studies and it is located 1,042 kilometers north-northwest of since it gives accurate information regarding land Buenos Aires. Santiago del Estero is placed in a change across time and can be calculated quickly transition zone between the temperate climates from satellite data (XI ET AL., 2019) and, unlike of the Pampas, and the subtropical climates of the NDVI, which is dependent on and varies with Chaco region and according to the Köppen climatic conditions, these built-up indices remain classification, it has a hot semi-arid climate (BSh). The annual precipitation of 695 mm is concentrated occupy an entire block, out of the 98 blocks in the mostly from November to March with the occurrence neighborhood (Fig. 2). It is an area where concrete of frosts between May and August and the average predominates, with few green spaces and trees annual temperature is 21.5 °C (ROGER ET AL., 2016). (Fig. 3). It is not possible to know exactly the population Table 1. Increment of population and housing units (1991- of the downtown (central) neighborhood of the 2010) (Source: INDEC, 1991, 2001, and 2010 Census) city of Santiago del Estero. However, in the period Population in Santiago Housing Units in Year 1991-2010 (Table 1) the number of inhabitants del Estero city Capital Dep. residing in the city increased by 33%, ten more 189 947 42 241 percentage points than the national average. On the 230 614 53 711 other hand, housing also registered a significant 2001 increase in the Capital department (where the city of 252 192 78 281 Santiago del Estero is located and comprises 94% 1991-2010 32.77 [%] 85.32 [%] of the population) with an increment of 85% in the Variation same period. The central neighborhood of the city has a perimeter of 5.10 km and an area of 1.55 km . In its interior, 25 census blocks are located totally or partially (each of them contains approximately 200 houses) (Fig. 1). This area contains 1965 pixels which are sufficient to carry out a microscale study. An inventory of the street trees carried out by ARIAS & CELEMIN (20218), indicates how the vegetation is located, mainly to the central-south and south-east of the study area, as the abundance increases from the interior to the bordering avenues of the neighborhood. It also registered the location of green spaces, where only two Fig. 1. Study area location Fig. 2. Tree abundance and green spaces in the study area engine/datasets/catalog/LANDSAT_LT05_C01_T1_S R#description). On the GEE platform (https://earthengine.google.org/), all data processing was done using cloud computing technology. To this purpose, a script was created with the goal to acquire the overall value of NDVI and UI for the study area for all images available throughout the study period. Landsat 7 ETM+ images were not used due to the failure of the Scan Line Corrector (SLC) in 2003, resulting in some areas that are imaged twice and others that are not imaged at all. In addition, only cloudless images are selected through the script, for both satellite catalogues. Finally, the mean was obtained from the images of the entire study area using the ReduceMean feature. There are different indices obtained from satellite images that allow us to know the state of Fig. 3. Street photo in the study area (-27.7912; -64.2572), the vegetation where the Normalized Difference Arias, 20/04/2022 Vegetation Index (NDVI) is the most recognized and used. It is calculated by using the following Given the capacity of the GEE to create visual expression: information continuously, a video was made with the monthly evolution of the NDVI in the 1991- 2011 period, as a way to provide a better visual interpretation of the study area and of the spatial where NIR is near-infrared reflectivity (Landast 5 distribution of vegetation greenness. For this, a TM band 4) and R is reflectivity in red (Landsat 5 script was created with Landsat 5 TM 32-Day TM band 3). The image value is delimited by the NDVI Composite data. Video available at: range -1 and 1 and the closer it is to 1 the greater https://drive.google.com/file/d/1x31fiqykxCBin the presence of healthy vegetation in a place. beI0FefnYknnq9pMK_x/view?usp=sharing The UI normalizes the NIR and SWIR 2 band, which makes full use of the inverse relationship 3. Methods between the brightness of the NIR and SWIR in built-up area (XI ET AL., 2019). The GEE catalogue for the Landsat 5 TM was used in this investigation. This satellite was a low- Earth-orbit platform launched on March 1, 1984, and administered jointly by the USGS and the National Aeronautics and Space Administration where SWIR2 is Short Wave Infrared 2 (Landast (NASA). The USGS Earth Resources Science and 5 TM band 7) and NIR is near-infrared reflectivity Observation Center collected and distributed (Landast 5 TM band 4). Landsat 5 TM data (EROS). Landsat 5 TM was The entire data analysis of both indices was formally deactivated on June 5, 2013, after 29 conducted using GEE. An imported shapefile of the years in space. Central area of Santiago del Estero city was imported GEE provides online access to archived Landsat into the GEE for extracting the NDVI and UI values data, including Landsat 5 TM from 1991 to 2011. over a 19-year timelapse (1992–2011). With the We used the "LANDSAT/LT05/C01/T1_SR" catalogue resulting dataset, which provided no data for the that contains atmospherically corrected surface year 1991, we computed a mean value of the reflectance images from the Landsat 5 TM sensor. NDVI and UI from all pixels in the study area for These images contain 4 visible and near-infrared all the 60 images in order to globally know the (VNIR) bands and 2 short-wave infrared (SWIR) temporal evolution of both indices. Furthermore, bands processed to orthorectified surface reflectance, the seasonality patterns of the NDVI and UI were and one thermal infrared (TIR) band processed to examined using long-term monthly means. Next, orthorectified brightness temperature. The VNIR we studied the annual performance to examine and SWIR bands have a resolution of 30m / pixel the fluctuation of the NDVI and UI over time. (https://developers.google.com/earth- Finally, the relationship between both indices was Feb 5, 1992 to Sept 25, 2011. An initial visual analyzed using the R-squared procedure. interpretation of the original dataset allows us to observe a continuous decrease in the NDVI 4. Results while the UI records show a general increasing trend (Fig. 4). The correlation of both datasets 4.1. Dataset analysis registers a negative slope, with a high R-squared of -0.863 with an NDVI mean of 0.136 and 0.023 The data processing from the GEE platform for the UI. resulted in a total of 60 images corresponding to Fig. 4. NDVI-UI evolution (complete dataset) 4.2. Annual and temporal variations of NDVI and UI decreases a little. The NDVI has an annual mean for the 19-year period of 0.136, while the UI registers The next result models the trend of the NDVI a mean of 0.023. Furthermore, the NDVI has a higher and UI over 19 years in the central area of the city R-squared in the linear regression model with a of Santiago del Estero (Fig 5). By averaging the mean score of 0.66, while the UI has a 0.38 score. Both have NDVI and UI values over all pixels in the study area similar Standard Deviation values with 0.021 and the linear regression model was adopted to identify 0.029 respectively. the temporal variation of both indices. The mean The seasonal variability of NDVI and UI illustrates NDVI values vary from 0.161 in 1992 to 0.103 in the NDVI and UI seasonal patterns for the study 2011, while the mean UI has a score of 0.003 in 1992 area (Fig. 6). The seasonal peaks of the NDVI are and 0.036 in 2011. Consequently, since 1992, there observed during spring and summer seasons, has been an overall constant decaying trend in while the lowest NDVI mean scores are observed NDVI while the UI has the opposite output, although in the winter. On the contrary, the UI presents with more abrupt fluctuations. In the first years it negative values in the summer months and from registers negative values, with the exception of mid-autumn it increases considerably reaching 1992, that quickly rise to positive scores beginning in its peak in late winter (August-September). 1997. From that date on the UI plateaus and even Fig. 5. Annual performance of NDVI and UI Fig. 6. Seasonal variability of NDVI and UI 5. Discussion 5.1. Densification and loss of vegetation cover The impact of urbanization on natural ecosystems The current study employed Landsat 5 TM and on habitat quality is a topic of current study images to examine the NDVI and UI evolution for as the relationship between urban dynamics and the downtown neighbourhood of Santiago del plant communities involves processes with complex Estero from 1992 to 2011. In the context of characteristics. Population growth is decisive in densification in the research area, the results suggest the decline of NDVI, with strong negative correlations a decreasing presence of vegetation. Densification in urban sites, especially in Latin America (ARBOIT is an urbanization approach that entails increasing & MAGLIONE, 2018). A global scale study focused the amount of built space and creating compact on urban vegetation cover (RICHARDS & BELCHER, cities rather than expanding cities in order to 2019) shows how it had decreased in most urban make better use of limited space (EGGIMANN ET AL., areas between 2000 and 2015, mainly in less 2021). In Argentina, the analysis of several cities developed countries; however, vegetation cover in the Region of Cuyo (Argentina) carried out by slightly increased in some urban areas in eastern ARBOIT & MAGLIONE (2018) highlights population North America and parts of Europe. Urbanization growth as a factor in the decline of NDVI. Other that does not take care of its own natural landscape studies on the spatial and temporal analysis of NDVI poses a threat to the quality of the urban environment for the town of Monte Hermoso in the province of and, thus, the quality of life of the inhabitants Buenos Aires, Argentina, in the period 2008-2012 (YEPEZ ET AL., 2014). Remote sensing techniques can (FERRELLI ET AL., 2018) and for the city of Bahía be used to analyze the increase of impervious Blanca also shows a loss of NDVI values (FERRELLI surfaces and vegetation status to gain a better ET AL., 2015). More specifically related to densification understanding of urban areas. Spectral indices in cities of Argentina, we can mention the work of have the advantages of being easy to build, MERLOTTO ET AL., (2012) who carried out a study parameter-free, and useful in land surface information on land cover between 1967-1984 for the towns of extraction applications and provide geographical Quequén and Necochea, with results showing that and temporal data that are utilized to assist urban the process of densification of urban occupation populations and decision-makers in maintaining, is significantly higher than that of the expansion. or improving, their cities' quality of living in the On the other hand, a study focused on twelve future (HIDAYATI ET AL., 2021). The loss of vegetation cities in northern Argentina (the poorest region in a context of coexistence of new urban peripheries of the country) (PAOLINI ET AL., 2016) found that and densification in Latin American cities urges the dynamics of urban growth in this area were urban planning to enact strategies that not only dominated by patterns of expansion rather than facilitate the creation of new green spaces, but homogeneous densification, although both processes also to form an urban framework sustained by coexist to some extent, and which seems to be the regulations that favour the planting and maintenance case for the city of Santiago del Estero. For Latin of vegetation in private spaces (DE LA BARRERA & America, the work of VEGA ET AL. (2019) records a HENRÍQUEZ, 2017). considerable loss of vegetation cover by estimating NDVI for urban areas of the city of Iquitos (Peru) (MUGIRANEZA ET AL., 2020). The Landsat library in between the years 1999-2009. Another recent case GEE comprises more than three decades of Earth is in the city of Medellín, for the 1986-2016 timelapse observation photos, giving researchers a unique shows a greater loss of vegetation in the densest chance to track land cover changes through time area of the town (SOTO-ESTRADA, 2019). Similar results with high temporal and spatial resolution. Because are presented by DE CARVALHO & SZLAFSZTEIN (2018) of their great spatiotemporal resolution, Landsat- in a case study of the city of Belém (Brazil) for the based time series data are ideal for detecting period 1986-2009 using Landsat 5 TM images. vegetation change (HUANG ET AL., 2018). This study A limitation to appropriately measure demonstrated that a simple analysis of NDVI and UI densification in Argentina (either due to an trends in a local context can be easily replicated in increase in population and/or buildings) is the other areas. The scripts are stored in the GEE platform way in which the censuses in Argentina record and by replacing the polygons corresponding to a the information. Moreover, in most cases, the study area with another area of interest, the total population can be only approximately known information for a new set of data is generated. since the unit of measurement of the censuses Finally, the importance of obtaining temporally (census blocks) rarely coincides with the limits of continuous data is highlighted, unlike the census a neighborhood. For example, in the interior of information that is available on specific dates, the downtown neighborhood of the city of Santiago separated by many years. del Estero, there are 18 census blocks, but 7 partially occupy the surface of the neighborhood. 6. Conclusions Built-up indices are sensitive to construction and are frequently used to represent the degree Latin American cities experience simultaneous of development and density of a built-up area. processes of urban growth and densification. Both Different indices have been developed to determine are complementary but each has its own dynamics. built-up areas from satellite images, however, The study of densification, which can be observed none have obtained a much higher precision than in greater detail in the central areas of cities, is the rest, resulting in a proliferation of indices. not as advanced as that of urban growth, even They all have potential and limitations, while the more so it is linked with elements of the local NDBI and the UI stand out for their ease of landscape, such as vegetation. The development implementation. Previous research revealed that of artificial spaces and the replacement of natural built-up indices were better than NDVI for features are known to function as the main drivers of quantitatively detecting land changes over time (XIE change in the local space. In this context we found ET AL., 2021; KUMARI ET AL., 2020). However, this is that the NDVI showed a decreasing trend in the not the case in this study since the interpretation timelapse under consideration, while the UI of the graphs and the data indicates a slightly performance registered the opposite trend in a greater variability of the UI than that of the 19-year period in the central area of the city of vegetation index. In addition, the results arise Santiago del Estero (Argentina) which has rapidly new interrogations that will be addressed in increased its population and housing units in the future works. For example, it is interesting to last decades. Both indices also presented a high observe how the UI index plateaus since the late negative correlation. Because the results were nineties. This could suggest that the area has valuable and consistent, the approach can be reached some degree of building saturation, used in other cities. The loss of vegetation and the although this does not explain the decrease of increase in built area are common characteristics NDVI values since that date. of Latin American cities that require more attention, particularly in the context of climate 5.1. Potentialities and limitations of the use of GEE change. The findings of this study can help to implement local policies aimed at improving and Long-term satellite imaging is critical for increasing the area of green space in the study area. understanding dynamic land cover, and it is especially good for detecting vegetation changes. References A variety of sensors produce images with varying Andersson E., Haase D., Scheuer S., Wellmann T. 2020. resolutions with the goal of detecting specific types Neighbourhood character affects the spatial extent and of land cover. GEE is a cloud-based geospatial magnitude of the functional footprint of urban green processing platform that provides a vast collection of infrastructure. Landscape Ecology, 35, 7: 1605–1618. data for analyzing free satellite imagery, producing Arboit M.E., Maglione D.S. 2018. Multi-temporal and multi- spatial analysis of the normalized difference vegetation statistics and maps, and graphical representations index (NDVI) and soil-adjusted vegetation index (SAVI) in Firozjaei M.K., Sedighi A., Kiavarz M., Qureshi S., Haase D., forested urban centers and irrigated oases, with dry Alavipanah S.K. 2019. Automated built-up extraction climates. Boletín de Estudios geográficos, 109 [in Spanish]. index: A new technique for mapping surface built-up Arias M.E., Celemin J. 2021. Spatial distribution of street areas using LANDSAT 8 OLI imagery. Remote Sensing, 11, trees in the Center of the city of Santiago del Estero 17: 1966. (Argentina). Revista da Casa da Geografia de Sobral, 23: Fontenelle M.R., Lorente S., Gonçalves Bastos L.E. 2015. The 434 – 454 [in Spanish]. impact of urbanization on air flow pattern: the case of Bensús Talavera V. 2018. (Un)planned densification of a Rio de Janeiro. International Journal of Green Energy, 12: metropolis. The case of the Metropolitan Area of Lima 908–916. 2000–2014. Revista INVI, 33(92): 9–51 [in Spanish]. Gaw L.Y.F., Richards D.R. 2021. Development of spontaneous Berkowitz A.R., Nilon C.H., Hollweg K.S. 2003. The importance of vegetation on reclaimed land in Singapore measured by understanding urban ecosystems: Themes. [in:] A.R. NDVI. Plos one, 16, 1: e0245220. Berkowitz, C.H. Nilon, K.S. Hollweg (eds.) Understanding Grover A., Singh R.B. 2015. Analysis of urban heat island (UHI) urban ecosystems - A new frontier for science and education. in relation to normalized difference vegetation index Sringer-Verlag, New York: 15–17. (NDVI): A comparative study of Delhi and Mumbai. Berndtsson R., Becker P., Persson A., Aspegren H., Environments, 2, 2: 125–138. Haghighatafshar S., Jönsson K., ... Tussupova K. 2019. He C., Shi P., Xie D., Zhao Y. 2010. Improving the Normalized Drivers of changing urban flood risk: A framework for Difference Built-Up Index to Map Urban Built-Up Areas action. Journal of Environmental Management, 240: 47–56. Using a Semiautomatic Segmentation Approach. Remote Cellucci C., Sivo M.D. 2021. Green Densification Strategies in Sensing Letters, 1: 213–221. Inner City for Psycho-Physical-Social Wellbeing. [in:] T. Hidayati I.N., Kusumawardani K.P., Ayudyanti A.G., Prabaswara Ahram, R. Taiar, F. Groff (eds) Human Interaction, R.R. 2021. Urban Biophysical Quality Modelling Based on Emerging Technologies and Future Applications IV. IHIET- Remote Sensing Data in Semarang, Indonesia. Geography, AI 2021. Advances in Intelligent Systems and Computing, Environment, Sustainability, 14, 3: 14–23. vol 1378, Springer, Cham: 350–358. Huang S.L., Wang S.H., Budd W.W. 2009. Sprawl in Taipei’s Coppola E. 2012. Densification vs Urban Sprawl. Tema-Journal peri-urban zone: Responses to spatial planning and of Land Use Mobility And Environment, 5, 1: 131–143. implications for adapting global environmental change. Darchen S., Poitras C. 2018. Accommodating densification Landscape and Urban Planning, 90: 20–32. and social sustainability in the inner city: Case study of INDEC. 1991 Census. Argentina. Griffintown, Montreal. [in:] G. Searle (ed.) Compulsory INDEC. 2001 Census. REDATAM Database, Argentina. Property Acquisition for Urban Densification. Routledge, INDEC. 2010 Census. REDATAM Database, Argentina. London: 67–80. Jianya G., Haigang S., Guorui M., Qiming Z.A. 2008. Review of De Carvalho R.M., Szlafsztein C.F. 2018. Urban vegetation Multi-Temporal Remote Sensing Data Change Detection loss and ecosystem services: The influence on climate Algorithms. The International Archives of the Photo- regulation and noise and air pollution. Environmental grammetry. Remote Sensing and Spatial Information Sciences, Pollution, 245: 844–852. 37: 757–762. De la Barrera F., Henríquez C. 2017. Vegetation cover change Kawamura M., Jayamana S., Tsujiko Y. 1996. Relation in growing urban agglomerations in Chile. Ecological between Social and Environmental Conditions in Colombo Indicators, 81: 265–273. Sri Lanka and the Urban Index Estimated by Satellite del Río M.V. 2017. Impact of intensive residential densification Remote Sensing Data. International Archieve of Photo- on the segmentation of the urban fabric of Santiago: a grammetry and Remote Sensing, 31 (B7): 321–326. quantitative approach. Revista, 180(40) [in Spanish]. Kumari B., Tayyab M., Ahmed I.A., Baig M.R.I., Khan M.F., Do J., Ahn S., Kang J. 2021. Urbanization effect of mega Rahman A. 2020. Longitudinal study of land surface sporting events using sentinel-2 satellite images: The case temperature (LST) using mono-and split-window algorithms of the Pyeongchang olympics. Sustainable Cities and and its relationship with ND VI and NDBI over selected Society, 74: 103158. metro cities of India. Arabian Journal of Geosciences, 13(19): Eggimann S., Wagner M., Ho Y.N., Züger M., Schneider U., 1–19. Orehounig K. 2021. Geospatial simulation of urban Lemonsu A., Viguié V., Daniel M., Masson V. 2015. Vulnerability to neighbourhood densification potentials. Sustainable Cities heat waves: impact of urban expansion scenarios on and Society, 72: 103068. urban heat island and heat stress in Paris (France). Urban Emilsson T., Ode Sang Å. 2017. Impacts of climate change on Climate, 14: 586–605. urban areas and nature-based solutions for adaptation. Li J., Roy D.P. 2017. A global analysis of Sentinel-2A, Sentinel- [in:] N. Kabisch, H. Korn, J. Stadler, A. Bonn (eds) Nature- 2B and Landsat-8 data revisit intervals and implications Based Solutions to Climate Change Adaptation in Urban for terrestrial monitoring. Remote Sensing, 9,9: 902. Areas. Theory and Practice of Urban Sustainability Transitions. Li X., Sunikka-Blank M. 2021. Urban densification and social Springer, Cham 15–27. capital: neighbourhood restructuring in Jinan, China. Ferrelli F., Bustos M.L., Huamantinco Cisneros M.A., Piccolo Buildings and Cities, 2, 1: 244–263. M.C. 2015. Use of satellite images for the study of the Libertun de Duren N., Guerrero Compeán R. 2016. Growing thermal distribution in different land covers of the city of resources for growing cities: Density and the cost of Bahía Blanca. Revista de Teledetección, 44: 31–42 [in Spanish]. municipal public services in Latin America. Urban Studies, Ferrelli F., Cisneros M.A.H., Delgado A.L., Piccolo M.C. 2018. 53, 14: 3082–3107. Spatial and temporal analysis of the LST-NDVI relationship Mattos C. 2016. Lógica financiera, geografía de la financiarización for the study of land cover changes and their contribution to y crecimiento urbano mercantilizado. [in:] F. Link, J. urban planning in Monte Hermoso, Argentina. Documents Noyola y A. Orellana (eds.), Urbanización planetaria y la d’Anàlisi Geogràfica, 64, 1: 25–47. reconstrucción de la ciudad. RIL Editores, Santiago de Ferro J.S. 2001. Expansion or Densification? Reflections on Chile, 29–55 [in Spanish]. the Bogotá Case. Bitácora Urbano-Territorial, 5, 1: 21–35 Merlotto A., Piccolo M.C., Bértola G.R. 2012. Urban growth [in Spanish]. and land use/cover changes in the cities of Necochea and Quequén, Buenos Aires, Argentina. Revista de Geografía Trombetti M., Riaño D., Rubio M.A., Cheng Y.B., Ustin S.L. Norte Grande, 53: 159–176 [in Spanish]. 2008. Multi-temporal vegetation canopy water content Mugiraneza T., Nascetti A., Ban Y. 2020. Continuous retrieval and interpretation using artificial neural monitoring of urban land cover change trajectories with networks for the continental USA. Remote Sensing of landsat time series and landtrendr-google earth engine Environment, 112 (1): 203–-215. cloud computing. Remote Sensing, 12, 18: 2883. UN - Habitat (United Nations Human Settlements Programme). Næss P., Saglie I.L., Richardson T. 2020. Urban sustainability: 2012. Leveragig Density: Urban Patterns for a Green is densification sufficient? European Planning Studies, 28, Economy. UN Habitat, Nairobi. 1: 146–165. Wang S.H., Huang S.L., Huang P.J. 2018. Can spatial planning Nolè G., Lasaponara R., Murgante B. 2013. Applying spatial really mitigate carbon dioxide emissions in urban areas? autocorrelation techniques to multi-temporal satellite A case study in Taipei, Taiwan. Landscape and Urban data for measuring urban sprawl. International Journal of Planning, 169: 22–36. Environmental Protection, 3, 7: 11. Vargas-Bolaños C., Orozco-Montoya R., Vargas-Hernández A., Paolini L., Aráoz E., Gioia A., Powell P.A. 2016. Vegetation Aguilar-Arias J. 2020. Methodology for determining the productivity trends in response to urban dynamics. growth of the urban sprawl in the capitals of the Central Urban Forestry & Urban Greening, 17: 211–216. American region (1975–1995–2014). Revista Geográfica Pauleit S., Ennos R., Golding Y. 2005. Modeling the environmental de América Central, 64: 41–74 [in Spanish]. impacts of urban land use and land cover change–a study Vega J.J.P., Zárate-Gómez R., Vela R.J.M., Brañas M.M., Rios in Merseyside, UK. Landscape and Urban Planning, 71, 2– J.E.B. 2019. (Predicción de la pérdida de la cobertura 4: 295–310. vegetal por aumento de áreas urbanas en Iquitos, Perú). Richards D.R., Belcher R.N. 2019. Global changes in urban Ciencia Amazónica (Iquitos), 7, 1: 37–50 [in Spanish]. vegetation cover. Remote Sensing, 12 (1): 23. Xi Y., Thinh N.X., Li C. 2019. Preliminary comparative Roger E., Palacio M., Coria O., Díaz R. 2016. Notes on the assessment of various spectral indices for built-up land urban flora cultivated in the city of Santiago del Estero, derived from Landsat-8 OLI and Sentinel-2A MSI imageries. Argentina. Multequina, 25, 1: 29–41 [in Spanish]. European Journal of Remote Sensing, 52, 1: 240–252. Sinha P., Verma N.K., Ayele E. 2016. Urban built-up area Xie Q., Sun Q. 2021. Monitoring the Spatial Variation of extraction and change detection of Adama municipal area Aerosol Optical Depth and Its Correlation with Land using time-series Landsat images. International Journal Use/Land Cover in Wuhan, China: A Perspective of Urban of Advanced Remote Sensing and GIS, 5, 8: 1886–1895. Planning. International Journal of Environmental Research Skrede J., Berg S.K. 2019. Cultural heritage and sustainable and Public Health, 18, 3: 1132. development: the case of urban densification. The Historic Xu H. 2008. A New Index for Delineating Built-Up Land Environment: Policy & Practice, 10, 1: 83–102. Features in Satellite Imagery. International Journal of Soto-Estrada E. 2019. Estimation of the urban heat island in Remote Sensing, 29: 4269–4276. Medellin, Colombia. Revista Internacional de Contaminación Yépez Rincón F.D., Lozano García D.F. 2014. Mapping of Ambiental, 35, 2: 421–434. urban trees with aerial. Revista Mexicana de Ciencias Tillie N., Borsboom-van Beurden J., Doepel D., Aarts M. 2018. Forestales, 5, 26: 58–75 [in Spanish]. Exploring a stakeholder based urban densification and Zhao H.M., Chen X.L. 2005. Use of Normalized Difference greening agenda for Rotterdam inner city—accelerating Bareness Index in Quickly Mapping Bare Areas from the transition to a liveable low carbon city. Sustainability, TM/ETM+. Geoscience and Remote Sensing Symposium, 10(6), 1927. 3(25–29): 1666–1668. Treija S., Bratuškins U., Koroļova A. 2018. Urban Densification of Zurqani H.A., Post C.J., Mikhailova E.A., Allen J.S. 2019. Large Housing Estates in the Context of Privatisation of Mapping urbanization trends in a forested landscape Public Open Space: the Case of Imanta, Riga. Architecture using Google Earth Engine. Remote Sensing in Earth Systems & Urban Planning, 14, 1: 105 – 110. Sciences, 2, 4: 173–182.
Environmental & Socio-economic Studies – de Gruyter
Published: Sep 1, 2022
Keywords: vegetation cover loss; densification; NDVI; Google Earth Engine; Santiago del Estero; Argentina
Access the full text.
Sign up today, get DeepDyve free for 14 days.