Evaluating Multi-Sensor Nighttime Earth Observation Data for ... - MDPI

remote sensing Article

Evaluating Multi-Sensor Nighttime Earth Observation Data for Identification of Mixed vs. Residential Use in Urban Areas Christoph Aubrecht 1, *,† and José Antonio León Torres 2,† 1 2

* †

Social, Urban, Rural & Resilience (GSURR), The World Bank, Washington, DC 20006, USA Institute of Engineering, National Autonomous University of Mexico (UNAM), Mexico City 04510, Mexico; [email protected] Correspondence: [email protected] These authors contributed equally to this work.

Academic Editors: Ioannis Gitas and Prasad Thenkabail Received: 28 November 2015; Accepted: 25 January 2016; Published: 4 February 2016

Abstract: This paper introduces a novel top-down approach to geospatially identify and distinguish areas of mixed use from predominantly residential areas within urban agglomerations. Under the framework of the World Bank’s Central American Country Disaster Risk Profiles (CDRP) initiative, a disaggregated property stock exposure model has been developed as one of the key elements for disaster risk and loss estimation. Global spatial datasets are therefore used consistently to ensure wide-scale applicability and transferability. Residential and mixed use areas need to be identified in order to spatially link accordingly compiled property stock information. In the presented study, multi-sensor nighttime Earth Observation data and derivative products are evaluated as proxies to identify areas of peak human activity. Intense artificial night lighting in that context is associated with a high likelihood of commercial and/or industrial presence. Areas of low light intensity, in turn, can be considered more likely residential. Iterative intensity thresholding is tested for Cuenca City, Ecuador, in order to best match a given reference situation based on cadastral land use data. The results and findings are considered highly relevant for the CDRP initiative, but more generally underline the relevance of remote sensing data for top-down modeling approaches at a wide spatial scale. Keywords: top-down modeling; urban areas; nighttime lights; DMSP; VIIRS; human activity; residential use; mixed use; global spatial data; CDRP

1. Introduction Issues of urban development are increasingly being addressed at the global scale, with international non-governmental organizations (NGOs) and development institutions often setting the path and moving the public agenda forward. Regularly published reports such as the United Nations’ World Urbanization Prospects [1] or the World Bank’s World Development [2] and Global Monitoring Reports [3] address fundamental issues and define key research questions to be tackled by the scientific and international development community. In that context it has become more and more evident that spatial data is playing a crucial role for consistent cross-regional analyses and unbiased evaluation of locally implemented actions. Remote sensing data in particular provide a rich and globally consistent source for analyses at multiple levels. At the global scale, different aspects have to be considered than for local-level spatial analyses, including consistency, scalability, retraceability etc. Several global project initiatives address these issues in various thematic domains. The World Bank’s Global Urban Growth Data Initiative, for example, addresses pending issues of regional definition and data incompatibilities and supports the international collaborative setup and development of

Remote Sens. 2016, 8, 114; doi:10.3390/rs8020114

www.mdpi.com/journal/remotesensing

Remote Sens. 2016, 8, 114

2 of 19

a consistent data set of global urban extents and associated population distribution patterns. In the same context, the Global Human Settlement Working Group, established under the umbrella of the Group on Earth Observations (GEO) (www.earthobservations.org/ghs), aims at establishing a new generation of global settlement measurements and products based on consistent high-resolution satellite imagery analysis. The presented study has been carried out within the framework of the World Bank’s Country Disaster Risk Profiles (CDRP) project initiative which has been successfully implemented at the continental scale for Central America [4] and is currently being expanded to the Caribbean Region. With the clear aim at extending to other regions, global applicability and easy transferability are considered crucial for the model setup. Global spatial datasets are therefore used throughout the CDRP project, with the presented approach specifically developed to support implementation of a disaggregated property stock exposure model, one of the key elements for subsequent disaster risk and loss estimation. While focusing primarily on natural hazards and risks, urban-rural identification and intra-urban classification aspects are highly relevant for setting the basic spatial framework for analysis [5]. This paper introduces a novel approach to geospatially identify and distinguish areas of mixed use from predominantly residential areas within urban agglomerations. After initial urban-rural classification at a 1 km grid level, that urban mask needs to be classified in residential and mixed use areas in order to spatially link accordingly compiled property stock information (e.g., from global tabular databases such as PAGER-STR [6]). The distinct identification of urban residential and mixed use areas serves as crucial input to define inventory regions for subsequent exposure assessment. Impervious Surface Area (ISA) data [7] based on remotely sensed nighttime lights from the Defense Meteorological Satellite Program’s (DMSP) OLS sensor (Operational Linescan System) are used as proxy to identify areas of peak human activity, often associated with a high likelihood of commercial and/or industrial presence. ISA is chosen due to its inherent correlation with built-up area (providing an indication on the percentage of built-up per grid cell) and thus good suitability for building stock related land use classification. Several ISA thresholds are tested for a case study in Cuenca City, Ecuador, in order to best match a given reference situation on the ground, where local-level cadastral land use data is used to identify the actual distribution ratio of residential vs. mixed use areas. Furthermore, unaltered nightlight intensity data as provided by the VIIRS sensor (Visible Infrared Imaging Radiometer Suite) [8] are evaluated as alternative to the ISA data. With the DMSP program fading out VIIRS provides the option for successive nighttime Earth Observation analyses due to its low light imaging capability. We apply the same methodological steps as for the ISA data in order to determine best-matching thresholds for binary land use classification and subsequently perform a comparative analysis of the results. Also scale effects are accounted for in that regard with VIIRS featuring a higher spatial resolution than OLS-based data products. Preliminary results of this study were presented at the ECRS-1 conference [9]. Extensive further research and integration of alternative data sources then lead to the multi-sensor approach illustrated in this paper, highlighting the relevance of global remote sensing data for top-down modeling approaches at wide spatial scale. Outcome is considered relevant for global urban spatial modeling in a variety of topical domains including urban monitoring, disaster risk management, and regional development. 2. Study Area and Data Due to availability of detailed in situ reference data for comparative analysis and evaluation of the proposed methodology, the city of Cuenca, Ecuador, was chosen as study area. Cuenca City is located in the mountainous southern region of Ecuador at an elevation of around 2500 m above sea level and is the capital of the Azuay province (Figure 1). The city stretches across an area of roughly 70 km2 and had an urban population of 329,928 inhabitants according to the census 2010. Latest figures of the Ecuador National Statistical Office (INEC) estimate an urban population of approximately 400,000 in 2015.

Remote Sens. 2016, 8, 114 Remote Sens. 2016, 8, 114

3 of 19 3 of 19

Study area Cuenca City, Ecuador. Figure 1. Study Ecuador.

The use use of of satellite-observed satellite-observed nighttime nighttime lights lights has has aa long long tradition tradition in in research research dealing dealing with with The monitoring urban areas and patterns of human and economic activity [10–16] as well as its impact on monitoring urban areas and patterns of human and economic activity [10–16] as well as its impact on the environment [17–19]. As opposed to attempts of using nighttime lights for basic delineation of the environment [17–19]. As opposed to attempts of using nighttime lights for basic delineation of urban areas areas or or as as weights weights for for population population disaggregation, disaggregation, in in this this paper paper we we rather rather aim aim at at exploring exploring the the urban use and value in determining intra-urban characteristics. use and value in determining intra-urban characteristics. Public-domain applications been restricted restricted to to one one Public-domain applications of of nighttime nighttime Earth Earth Observation Observation have have long long been satellite sensor, namely the Operational Linescan system (OLS) onboard the Defense Meteorological satellite sensor, namely the Operational Linescan system (OLS) onboard the Defense Meteorological Satellite Program Program (DMSP) (DMSP) platform platform [20]. [20]. More Visible Infrared Infrared Imaging Imaging Satellite More recently, recently, data data from from the the Visible Radiometer Suite Suite (VIIRS) (VIIRS) sensor sensor onboard onboard the the Suomi Suomi NPP NPP satellite satellite platform platform have have become become available, Radiometer available, providing both higher spatial and radiometric resolution and being considered as natural successor successor providing both higher spatial and radiometric resolution and being considered as natural to the fading-out DMSP-OLS series [21]. The commercial satellite EROS-B also offers nighttime to the fading-out DMSP-OLS series [21]. The commercial satellite EROS-B also offers nighttime acquisition capability, capability,even even at very high spatial resolution [22]. However, global-scale and acquisition at very high spatial resolution [22]. However, global-scale and temporally temporally continuous open data availability remains restricted to DMSP-OLS and NPP-VIIRS, continuous open data availability remains restricted to DMSP-OLS and NPP-VIIRS, therefore the only therefore the onlygiven reasonable choice given the scope ofCDRP the above-outlined CDRP initiative. In the reasonable choice the scope of the above-outlined initiative. In the following, we briefly following,the wetwo briefly the two sourcesCity we use the (1) Cuenca City case study, the introduce dataintroduce sources we use fordata the Cuenca casefor study, the Impervious Surface(1)Area Impervious Surface Area (ISA) product derived from DMSP-OLS; and (2) VIIRS Day/Night band (ISA) product derived from DMSP-OLS; and (2) VIIRS Day/Night band light intensity data. light intensity data. 2.1. Impervious Surface Area (ISA) Data, Derived from DMSP-OLS 2.1. Impervious Surface Area (ISA) Data, Derived from DMSP-OLS The OLS sensor onboard the DMSP satellite series is able to detect faint light on the Earth’s TheatOLS the DMSPinsatellite series is able to detect faint designed light on the Earth’s surface nightsensor due toonboard its high sensitivity the visible spectrum. While initially to monitor surface at night due to its high sensitivity in the visible spectrum. While initially designed to monitor cloud coverage, that low light imaging capacity allows identification of various light emitting sources cloud coverage, low light capacity of various light emitting including human that settlements andimaging associated human allows activityidentification patterns [20]. The National Geophysical sources including human settlements and associated human activity patterns [20]. The National Data Center (NGDC) of the National Oceanic and Atmospheric Administration (NOAA) is processing Geophysical Center (NGDC) of the National Oceanic Atmospheric Administration and archivingData OLS imagery, thereby also making certain derivedand products accessible to the public. (NOAA) is data processing and archiving OLS imagery, thereby also area making certain products DMSP-OLS was first used to approximate impervious surface (ISA) in the derived early 2000s in the accessible to the public. DMSP-OLS data was first used to approximate impervious surface area development of a national-scale model for the conterminous United States [23]. The ISA approach (ISA) in the early 2000s in the development of a national-scale model for the conterminous United was then consequently adjusted to global scale whereby a radiance-calibrated annual composite of States [23]. Theis ISA approach was then consequently adjusted to global scale nighttime lights analyzed in conjunction with ancillary data such as population counts.whereby Output isa radiance-calibrated annual composite of nighttime lights is analyzed in conjunction with ancillary consistently provided at 30 arc-sec spatial resolution giving indication on the distribution of manmade data such as population counts. is consistently provided at 30 arc-sec spatial resolution surfaces including buildings, roadsOutput and related elements [7].

3


4 of 19

giving indication on the distribution of manmade surfaces including buildings, roads and related 4 of 19 elements [7]. Due to its relevance for a broad set of applications, global impervious surface or general Duearea to itsmapping relevancehas forbeen a broad setfocus of applications, impervious surface ordifferent general built-up built-up in the of attentionglobal for a while with data from satellite area mapping has been in the focus of attention for a while with data from different satellite sensors used and various approaches implemented (overview provided by [24]), recentsensors efforts used and various approaches implemented provided by [24]), recent including high including high resolution products such as(overview the Global Urban Footprint (GUF)efforts [25] and the Global resolution products such the Global Footprint (GUF) [25]data andset theisGlobal Human Settlement Human Settlement Layeras(GHSL) [26].Urban The DMSP-derived ISA unique in a sense that it Layer (GHSL) [26].extract The DMSP-derived ISA data set is unique a sense that it does directly does not directly built-up from satellite imagery butinuses artificial nightnot lighting asextract proxy built-up satellite imagerygeneral but uses artificial night lighting as proxy While a detected measure.from While a detected correlation between night lights measure. and impervious surfaces general correlation night ISA lightsproduct, and impervious basis for the global ISA provides the basis between for the global inherent surfaces patternsprovides point to the different human activities product, inherent patterns point to different human (e.g.,Given commercial, industrial) ratherofthan (e.g., commercial, industrial) rather than mere builtactivities structures. the scope and purpose the mere built study structures. Given therelation scope and purposewith of the presented study two-sided relation presented this two-sided to built-up a specific weight on this non-residential human to built-up with aisspecific weightrelevant. on non-residential human activity is set particularly relevant. activity patterns particularly Figure 2 shows the globalpatterns ISA data for the year 2010 Figure 2 shows globalCity ISA study data set for the year 2010 extracted for the Cuenca City study area. extracted for thethe Cuenca area. Remote Sens. 2016, 8, 114

Figure 2. Impervious (ISA)—30 arc-sec arc-sec grid grid for for Cuenca Cuenca City. City. Figure 2. Impervious Surface Surface Area Area (ISA)—30

2.2. VIIRS Day/Night Band Data 2.2. VIIRS Day/Night Band Data Since 2011 the VIIRS sensor onboard the Suomi NPP satellite platform provides a natural Since 2011 the VIIRS sensor onboard the Suomi NPP satellite platform provides a natural successor successor to DMSP-OLS with its panchromatic Day/Night Band (DNB) detecting dim nighttime to DMSP-OLS with its panchromatic Day/Night Band (DNB) detecting dim nighttime scenes in similar scenes in similar manner. Using advanced processing schemes (e.g., excluding/correcting data manner. Using advanced processing schemes (e.g., excluding/correcting data impacted by stray light), impacted by stray light), NOAA-NGDC is producing global monthly composite products featuring NOAA-NGDC is producing global monthly composite products featuring average radiance values at average radiance values at 15 arc-sec spatial resolution [27]. As supplementary product the number 15 arc-sec spatial resolution [27]. As supplementary product the number of cloud-free observations of cloud-free observations that was used to create the average composite is reported for each cell. that was used to create the average composite is reported for each cell. In addition to its superior spatial resolution, VIIRS offers a set of improvements over In addition to its superior spatial resolution, VIIRS offers a set of improvements over OLS-derived OLS-derived nighttime lights. These include lower light detection limits, improved dynamic range nighttime lights. These include lower light detection limits, improved dynamic range as well as as well as quantification and calibration options previously unavailable [21]. One of the few quantification and calibration options previously unavailable [21]. One of the few disadvantages of disadvantages of VIIRS, on the other hand, refers to the later overpass time (after midnight), when VIIRS, on the other hand, refers to the later overpass time (after midnight), when outdoor lighting is at outdoor lighting is at a significantly lower level as compared to early evening when OLS a significantly lower level as compared to early evening when OLS acquisitions are made. Figure 3 acquisitions are made. Figure 3 illustrates the VIIRS DNB data for the June 2015 monthly composite illustrates the VIIRS DNB data for the June 2015 monthly composite extracted for the Cuenca City extracted for the Cuenca City study area (left). For comparative purposes, we also aggregate that study area (left). For comparative purposes, we also aggregate that data to a 30 arc-sec grid (right), data to a 30 arc-sec grid (right), thus matching the resolution of the ISA product. thus matching the resolution of the ISA product. 2.3. Cadastral Cadastral Data Data 2.3. Since 2010 2010 the theMunicipality MunicipalityofofCuenca Cuencahas hasintensified intensified efforts collect specific information Since thethe efforts to to collect specific information on on all buildings located in the urban area of Cuenca City enabling the construction of a complete all buildings located in the urban area of Cuenca City enabling the construction of a complete cadastral cadastral including database including geo-localization and information oncharacteristics. building characteristics. database geo-localization and information on building 4

Remote 2016, 8, Remote Sens. Sens. 2016, 8, 114 Remote Sens. 2016, 8, 114

55of of19 19 5 of 19

Figure 3. VIIRS Day/Night Band (DNB) data for Cuenca City. 15 arc-sec grid (left). Aggregated 30 Figure (DNB) data forfor Cuenca City. 15 15 arc-sec grid (left). Aggregated 30 Figure 3. 3. VIIRS VIIRSDay/Night Day/NightBand Band (DNB) data Cuenca City. arc-sec grid (left). Aggregated arc-sec grid (right). arc-sec gridgrid (right). 30 arc-sec (right).

More specifically, this cadastral database contains detailed information on the use of each More specifically, this cadastral database contains detailed information on the use of each Moreallowing specifically, cadastral contains detailed information on the useoccupancy of each building, building, forthis example thedatabase distinction of residential and nonresidential types. building, allowing for example the distinction of residential and nonresidential occupancy types. allowing for example theas distinction and nonresidential types. Data sources Data sources that served input for of theresidential cadastral data base originate occupancy from different national entities Data sources that served as input for the cadastral data base originate from different national entities that input for theofcadastral originate fromof different national entities(INEC), such as the suchserved as theasMunicipality Cuenca,data the base National Institute Statistics and Census the such as the Municipality of Cuenca, the National Institute of Statistics and Census (INEC), the Municipality of Cuenca,Water the National InstituteService of Statistics and Census (INEC), (ETAPA), the Telecommunications, Telecommunications, and Sewage Company of Cuenca the National Telecommunications, Water and Sewage Service Company of Cuenca (ETAPA), the National Water and Sewage Service Company(SNGR) of Cuenca (ETAPA), the National Secretariat for Risk Management Secretariat for Risk Management and the University of Cuenca. After a validation and Secretariat for Risk Management (SNGR) and the University of Cuenca. After a validation and (SNGR) and the University of Cuenca. After a validation and filtering process, the cadastral building filtering process, the cadastral building data base eventually comprises 65,436 records [28]. Each filtering process, the cadastral building data base eventually comprises 65,436 records [28]. 2 Each data basefootprint eventually comprises 65,436 records Each building footprint record area is georeferenced building record is georeferenced and[28]. includes information on built-up (in m ) and building footprint record is georeferenced and2 includes information on built-up area (in m2) and 2 and includes information on built-up area (in m ) and occupancy type (Figure 4). Residential buildings occupancy type (Figure 4). Residential buildings thereby cover an area of 12.9 km2 , complemented occupancy type (Figure 4). Residential buildings thereby cover an area of 12.9 km , complemented 2 2 2 thereby an area of 12.9built-up km , complemented by a 4.3 km non-residential built-up area. by a 4.3 cover km non-residential area. by a 4.3 km2 non-residential built-up area.

Figure 4. 4. Cadastral Cadastral building footprints of Cuenca City, classified in in residential residential and and non-residential non-residential Figure buildingfootprints footprintsof of Cuenca CuencaCity, City, classified classified Figure 4. Cadastral building in residential and non-residential occupancy types. occupancy types. types. occupancy

For comparative purposes, we aggregate building footprint data to a 15 arc-sec and a 30 arc-sec For For comparative comparative purposes, purposes,we weaggregate aggregatebuilding buildingfootprint footprintdata datato toaa15 15arc-sec arc-secand andaa 30 30 arc-sec arc-sec grid respectively, thus matching the resolutions of the two analyzed nighttime lights data sets. grid respectively,thus thusmatching matching resolutions of two the analyzed two analyzed nighttime lights data sets.5 grid respectively, thethe resolutions of the nighttime lights data sets. Figure Figure 5 illustrates the aggregated grids for the non-residential share of the built-up area. On top, Figure 5 illustrates the aggregated grids for the non-residential share ofarea. the built-up area. On top, illustrates the aggregated grids for the non-residential share of the built-up On top, non-residential non-residential built-up percentage is shown. At the bottom, each cell’s contribution to the total non-residential built-up percentage is shown. At the bottom, each cell’s contribution to the total built-up area is visualized, whereby a main cluster in the center of the city is clearly depicted. built-up area is visualized, whereby a main cluster in the center of the city is clearly depicted. 5 5


6 of 19

built-up percentage is shown. At the bottom, each cell’s contribution to the total built-up area is visualized, whereby Remote Sens. 2016, 8, 114 a main cluster in the center of the city is clearly depicted. 6 of 19

Figure 5. Cadastral data for Cuenca City aggregated to grids at 15 arc-sec (left) and 30 arc-sec (right) Figure 5. Cadastral data for Cuenca City aggregated to grids at 15 arc-sec (left) and 30 arc-sec (right) resolution. Non-residential built-up percentage (top). Percent-contribution of each cell to the total resolution. Non-residential built-up percentage (top). Percent-contribution of each cell to the total non-residential built-up built-up area area (bottom). (bottom). non-residential

3. Methods 3. Methods As outlined in the introduction, the presented study was carried out under the framework of As outlined in the introduction, the presented study was carried out under the framework of the World Bank’s Central American Country Disaster Risk Profiles (CDRP) Initiative [4]. With that the World Bank’s Central American Country Disaster Risk Profiles (CDRP) Initiative [4]. With that kind of continental and global models, the implemented scale level plays an important role in kind of continental and global models, the implemented scale level plays an important role in defining defining the basic spatial units of analysis. Working on a 30 arc-sec resolution grid level (i.e., the basic spatial units of analysis. Working on a 30 arc-sec resolution grid level (i.e., approximately approximately 1 km at the equator)—frequently used for global models—the spatial identification 1 km at the equator)—frequently used for global models—the spatial identification and distinction and distinction of unique inventory regions is often not unambiguously possible at the grid cell level of unique inventory regions is often not unambiguously possible at the grid cell level due to the due to the well-studied mixed pixel issue [29,30]. While large urban residential areas as well as well-studied mixed pixel issue [29,30]. While large urban residential areas as well as certain dedicated certain dedicated industrial zones are still often built in rather compact manner and can thus indeed industrial zones are still often built in rather compact manner and can thus indeed cover entire grid cover entire grid cells, particularly commercial areas are commonly intertwined with residences cells, particularly commercial areas are commonly intertwined with residences forming wider areas of forming wider areas of mixed use. In order to appropriately identify urban non-residential areas in a mixed use. In order to appropriately identify urban non-residential areas in a spatial top-down model spatial top-down model it is therefore considered reasonable to assume a certain share of residential it is therefore considered reasonable to assume a certain share of residential occupancy throughout occupancy throughout and consider grid cells that also include a non-residential share as areas of and consider grid cells that also include a non-residential share as areas of mixed use. mixed use. For identification of those built-up urban areas that also feature a share of non-residential use, For identification of those built-up urban areas that also feature a share of non-residential use, we refer to the above-outlined nighttime Earth Observation data and derivative products as proxy we refer to the above-outlined nighttime Earth Observation data and derivative products as proxy measure. The assumption hereby is that intense lighting in that context is associated with a high measure. The assumption hereby is that intense lighting in that context is associated with a high likelihood of commercial and/or industrial presence, commonly clustered in certain parts of a city likelihood of commercial and/or industrial presence, commonly clustered in certain parts of a city (such as central business districts and/or peripheral commercial zones). Areas of low light intensity, in (such as central business districts and/or peripheral commercial zones). Areas of low light intensity, turn, can be considered more likely residential. in turn, can be considered more likely residential. The main objective of this study is thus to identify the light intensity thresholds that match best The main objective of this study is thus to identify the light intensity thresholds that match best the separated distribution of residential vs. mixed use areas on the ground. DMSP-OLS derived the separated distribution of residential vs. mixed use areas on the ground. DMSP-OLS derived ISA data and VIIRS-DNB data are both evaluated and comparatively analyzed for that purpose. It should be noted that the presented approach is proposed only for pre-identified urban areas [31], as for rural regions coarse-scale lighting intensity has reduced spatial correlation with built-up and other additional aspects come into play.


7 of 19

ISA data and VIIRS-DNB data are both evaluated and comparatively analyzed for that purpose. It should be noted that the presented approach is proposed only for pre-identified urban areas [31], as for rural regions coarse-scale lighting intensity has reduced spatial correlation with built-up and other additional aspects come into play. Referring to the Cuenca City cadaster data we distinguish purely residential areas from areas of mixed use. Using the building footprint area data we obtain that 75% of the total built-up area of Cuenca City features residential occupancy, complemented by 25% non-residential occupancy. At the aggregated 15 arc-sec level, the top 25% mixed use cells (covering 25 km2 out of the total 98.75 km2 ) account for 92% of the city’s total non-residential built-up area. We then use this bottom-up-determined distribution ratio to identify the appropriate lighting intensity thresholds in the top-down model. In order to define the relevant data value histograms for the threshold identification, we select all cells of the respective ISA and VIIRS data sets that fall within the pre-defined urban test case area of Cuenca City. In the case of the ISA data, the min-max value range is thereby identified as 5.7–77.8. For the VIIRS data, the min-max value range is identified as 3.3–73.8. In order to factor out potential effects generated by the mere difference in spatial resolution between ISA and VIIRS data we aggregate the original 15 arc-sec VIIRS data to a 30 arc-sec grid, thus enabling direct spatial comparability to the ISA grid. We iteratively apply several threshold cut-off points in the identified value ranges and compare the resulting areas of relatively low and relatively high ISA and VIIRS values respectively to the aggregated cadastral data. The eventually selected final cut-off point is that threshold value that produces the best-matching output with regard to the 75:25 cadaster-based residential vs. mixed use area distribution ratio. 4. Results 4.1. Identification of Residential vs. Mixed Use Areas Using ISA Data Table 1 illustrates the various tested ISA threshold values and the corresponding building use distribution ratios as derived from comparative spatial overlay with the aggregated cadastral data at the 30 arc-sec grid level. ISA min-max range and respective threshold values are shown in the left part of the table, with the percentile column indicating the relative value distribution. In mathematical terms the percentile value is derived as follows: (Threshold–Min)/(Max–Min). Specifically, that means that for the highlighted ID 4 the ISA value of 42 indicates the median value (50th percentile) in the distribution histogram. Half of the values in the study area under consideration thus feature an ISA value lower than 42 and the other half feature a higher value. Table 1. ISA distribution thresholds and corresponding building use distribution ratios (grey indicating selected best-matching threshold). ISA Distribution

ID 1 2 3 4 5

Built-Up Area Distribution

Min

Threshold

Max

Percentile

Residential Use

Mixed Use

5.7 5.7 5.7 5.7 5.7

15 22 39 42 59

77.8 77.8 77.8 77.8 77.8

13% 23% 46% 50% 74%

32.00% 52.00% 70.00% 74.00% 96.00%

68.00% 48.00% 30.00% 26.00% 4.00%

Spatially overlaid on the aggregated cadastral building use density grid (at the 30 arc-sec level, comparable to the ISA grid), that results in a 74% residential ratio and a 26% mixed use share, thus best-matching the bottom-up-derived 75% residential share (ISA data is provided in integer numbers, thus making it impossible to exactly match that 75% target value). Figure 6 maps the binary land use classification (residential use vs. mixed use) for the 5 tested ISA thresholds respectively. The share of mixed use area decreases thereby corresponding to the higher ISA cut-off points.


8 of 19 8 of 19

Figure 6. 6. Binary Binary land land use use classification classification of of Cuenca Cuenca City City based based on on ISA ISA thresholds thresholds from from Table Table 1. 1. Blue Blue Figure indicates residential and orange mixed use. Table record IDs are indicated in the figure as 1–5. indicates residential and orange mixed use. Table record IDs are indicated in the figure as 1–5.

Having the building-level cadastral data at hand enables not only determination of the binary Having the building-level cadastral data at hand enables not only determination of the binary land use distribution ratio, but furthermore allows us to consequently evaluate the degree of spatial land use distribution ratio, but furthermore allows us to consequently evaluate the degree of spatial overlap as a measure of model output accuracy. Using the above-identified ISA threshold, thus best overlap as a measure of model output accuracy. Using the above-identified ISA threshold, thus best matching the relative distribution of the two occupancy types (residential and mixed use), 82.8% of matching the relative distribution of the two occupancy types (residential and mixed use), 82.8% of the total non-residential building stock of Cuenca City (3.6 of 4.3 km22) is indeed captured within the the total non-residential building stock of Cuenca City (3.6 of 4.3 km ) is indeed captured within the selected top-down-derived binary mixed use mask. selected top-down-derived binary mixed use mask. 4.2. Identification of Residential vs. Mixed Use Areas Using VIIRS Data in Original Spatial Resolution 4.2. Identification of Residential vs. Mixed Use Areas Using VIIRS Data in Original Spatial Resolution Applying the same approach of iterative thresholding as illustrated for the ISA data we use Applying the same approach of iterative thresholding as illustrated for the ISA data we use VIIRS VIIRS data to perform comparative analysis. As outlined above, VIIRS data is evaluated in that data to perform comparative analysis. As outlined above, VIIRS data is evaluated in that context first context first at its original resolution level and furthermore at aggregated level matching the ISA at its original resolution level and furthermore at aggregated level matching the ISA resolution in order resolution in order to guarantee direct spatial comparability and factor out potentially biased scale to guarantee direct spatial comparability and factor out potentially biased scale effects. effects. 4.2.1. Application of VIIRS Data in Original Spatial Resolution 4.2.1. Application of VIIRS Data in Original Spatial Resolution As shown above with the ISA data, Table 2 relates the various tested VIIRS threshold values to As shown above with the ISA data, 2 relatesratios. the various tested threshold valuesare to the corresponding cadastral building useTable distribution Several lightVIIRS intensity thresholds the corresponding cadastral building use distribution ratios.built-up Several area lightdistribution intensity thresholds are applied iteratively, approximating the occupancy-type-specific shares on the applied iteratively, approximating the occupancy-type-specific built-up area distribution shares on ground. Spatial overlay of the VIIRS data and the corresponding aggregated cadastral building use the ground. overlay the VIIRS datathat and the the53rd corresponding aggregated cadastral building density grid Spatial (at the 15 arc-secoflevel) indicates percentile threshold exactly matches the use density grid (at the 15 arc-sec level) indicates that the 53rd percentile threshold exactly matches land use distribution as derived from the cadastral information, i.e., 75% residential and 25% mixed the share. land use distribution asbinary derived from cadastral (residential information, i.e., residential 25% use Figure 7 maps the land use the classification use vs.75% mixed use) for 5and selected mixed use share. Figure The 7 maps theofbinary (residential use vs. mixed use) for 5 thresholds respectively. share mixedland use use areaclassification decreases thereby corresponding to the higher selected thresholds respectively. The share of mixed use area decreases thereby corresponding to the VIIRS cut-off points. higher VIIRS cut-off points. Table 2. VIIRS distribution thresholds (original 15 arc-sec grid) and corresponding building use Table 2. VIIRS distribution thresholds (original 15 arc-sec grid) and corresponding building use distribution ratios (grey indicating selected best-matching threshold, orange indicating 50% threshold distribution ratios (grey indicating selected best-matching threshold, orange indicating 50% for comparison). threshold for comparison). VIIRS Distribution Built-up Area Distribution ID VIIRS Distribution Built-up Area Distribution Min Threshold Max Percentile Residential Use Mixed Use ID Min Threshold Max Percentile Residential Use Mixed Use 1 3.2 16.0 73.8 18% 32.00% 68.00% 3.23.2 16.025.0 73.8 73.8 18% 31% 32.00%52.00% 68.00% 2 1 48.00% 3 2 29.00% 3.23.2 25.038.5 73.8 73.8 31% 50% 52.00%71.00% 48.00% 4 3.2 40.6 73.8 53% 75.00% 25.00% 3 3.2 38.5 73.8 50% 71.00% 29.00% 5 3.2 57.0 73.8 76% 96.00% 4.00% 4 3.2 40.6 73.8 53% 75.00% 25.00%

5

3.2

57.0

73.8

76%

8

96.00%

4.00%


9 of 19 9 of 19

Figure Figure 7. 7. Binary land use classification of Cuenca City based on thresholds from original VIIRS data (Table and orange mixed use.use. Table record IDs are in the in figure (Table 2). 2).Blue Blueindicates indicatesresidential residential and orange mixed Table record IDsindicated are indicated the as 1–5. as 1–5. figure

degree of spatial overlap between the binary classified VIIRS data and the Results regarding the degree correspondinglyaggregated aggregatedcadastral cadastralgrid grid indicate of total the total non-residential building correspondingly indicate thatthat 76%76% of the non-residential building stock 2 2 stock of Cuenca Cityof (3.27 of 4.3 kmcaptured ) are captured the selected top-down-derived use of Cuenca City (3.27 4.3 km ) are withinwithin the selected top-down-derived mixedmixed use mask mask (using the identified best-matching 53rd percentile threshold). (using the identified best-matching 53rd percentile threshold). 4.2.2. Application of VIIRS Data Aggregated to a 30 arc-sec arc-sec Grid Grid In order to perform comparative analysis at identical scale levels, VIIRS data is aggregated to a 30 arc-sec iterative threshold determination. Table Table 3 illustrates the various threshold arc-secgrid gridbefore before iterative threshold determination. 3 illustrates the tested various tested values from the aggregated VIIRS data and the corresponding cadastral building use distribution ratios. threshold values from the aggregated VIIRS data and the corresponding cadastral building use To guarantee ratios. direct comparability thecomparability ISA-based analysis, the ISA-based thresholds for the aggregated VIIRS distribution To guarantee with direct with the analysis, the thresholds data areaggregated applied in VIIRS such a data way are thatapplied the building usea distribution (shown the right part of for the in such way that theratios building use in distribution ratios Table 3) in arethe identical to the tests carried out before thecarried ISA data. (shown right part of Table 3) are identical to using the tests out before using the ISA data. Table Table 3.3. VIIRS VIIRS distribution distribution thresholds thresholds (aggregated (aggregated 30 30 arc-sec arc-sec grid) grid) and and corresponding corresponding building building use use distribution ratios (grey indicating selected best-matching threshold, orange indicating 50% threshold distribution ratios (grey indicating selected best-matching threshold, orange indicating 50% for comparison). threshold for comparison). ID ID

VIIRS Distribution VIIRS Distribution

Built-up Distribution Built-up AreaArea Distribution

ThresholdMax MaxPercentile PercentileResidential Residential Mixed Min Min Threshold Use Use Mixed UseUse

1 22 3 43 54

2.7 2.7 2.7 2.7 2.7 2.7 2.7 2.7 2.7

12.2 12.2 19.3 19.3 34.1 34.1 36.9 36.9 57.1

65.1 65.1 65.1 65.1 65.1 65.1 65.1 65.1 65.1

15%

55%

5

2.7

57.1

65.1

90%

27% 50%

15% 27% 50% 55% 90%

32.00% 32.00% 52.00% 52.00% 70.00% 70.00% 75.00% 96.00% 75.00%

68.00% 68.00% 48.00% 48.00% 30.00% 30.00% 25.00% 4.00% 25.00%

96.00%

4.00%

Figure 8 maps the binary land use classification (residential use vs. mixed use) for the 5 tested Figurein 8 the maps the binaryVIIRS land data use classification vs. mixed use) the of 5 tested thresholds aggregated respectively. (residential As with theuse previous tests, thefor share areas thresholds the aggregated VIIRS data corresponding respectively. Astowith the previous the share of areas with mixedin occupancy decreases thereby the higher cut-off tests, points. with Spatial mixed overlay occupancy decreases thereby corresponding to the higher cut-off points. of the aggregated VIIRS data and the corresponding cadastral building use density Spatial overlay thepercentile aggregated VIIRS data and the the corresponding building and use grid indicates that theof55th threshold best matches target valuecadastral of 75% residential density griduse indicates that the 55th percentile threshold the target 25% mixed shares (as derived from the in situ cadaster best data),matches thus slightly highervalue than of for75% the residential and 25% mixed use shares (as derived from the in situ cadaster data), thus slightly higher original-resolution VIIRS data. Evaluating the degree of spatial overlap between the aggregated VIIRS than and for the VIIRSgrid, data.we Evaluating the degree spatial overlap between the data the original-resolution corresponding cadastral detect that 79% of theoftotal non-residential building 2 ) is captured within aggregated VIIRS cadastral grid, we detect that 79%binary of themixed total stock of Cuenca Citydata (3.4 and of 4.3the kmcorresponding the selected top-down-derived non-residential building stock of Cuenca City (3.4 of 4.3 km2) is captured within the selected use mask. top-down-derived binary mixed use mask.

9


10 of 19 10 of 19

Figure 8. 8. Binary Binary land land use use classification classification of of Cuenca Cuenca City City based based on on thresholds thresholds from from aggregated aggregated VIIRS VIIRS Figure data (Table (Table 3). 3). Blue Blue indicates indicates residential residential and and orange orange mixed use. Table Table record record IDs IDs are are indicated indicated in in the the data figure as as 1–5. 1–5. figure

5. Discussion 5. Discussion The application application of of ISA ISA and and alternatively alternatively VIIRS VIIRS data data to to identify identify intra-urban intra-urban occupancy occupancy type type The distribution patterns patterns that that we we outline outline in in this this paper paper and and the the corresponding corresponding findings findings include include several several distribution interesting aspects for further discussion. In the following we highlight three relevant points at interesting aspects for further discussion. In the following we highlight three relevant points at different stages in the model setup. First, we provide some background information on selection different stages in the model setup. First, we provide some background information on selection criteria of of VIIRS VIIRS data. data. Then, Then, we we discuss discuss the the actual actual differences differences in in the the outcome outcome of of the the proposed proposed binary binary criteria land use classification approach when implemented using ISA vs. VIIRS data. Finally, we highlight land use classification approach when implemented using ISA vs. VIIRS data. Finally, we highlight the the impact proper spatial delineation the model outcome by applying a spatially impact that that proper urbanurban spatial delineation has onhas theon model outcome by applying a spatially shrunk shrunkmask urban for the Cuenca urban formask the Cuenca City testCity case.test case. 5.1. 5.1. VIIRS Data Selection Initial Initial nightlights nightlights data selection has a big big influence influence on on the the model model outcome outcome and and is is particularly particularly important important in aa sense sense that that VIIRS VIIRS data data is is provided provided by by NOAA-NGDC NOAA-NGDC as as basic basic monthly monthly average average light light intensity intensity composites composites whereas ISA data data comes comes as as aa fully-processed fully-processed product product derived derived from from annual annual DMSP-OLS and calibrated calibratedwith withancillary ancillarybuilt-up built-upreference referenceinformation. information.The The number DMSP-OLS composites and number of of cloud-free observations a crucial factor for producing average composites as excessive cloud-free observations is aiscrucial factor for producing average composites as excessive cloudcloud cover cover can obscure light-emitting sources on the ground. monthlyproducts productsfewer fewer observations observations are can obscure light-emitting sources on the ground. In In monthly are potentially composite grids as compared to yearly products and average values potentially available availabletotocompute compute composite grids as compared to yearly products and average values can therefore moreturn easily turn outskewed to be skewed and non-representative of extended can therefore more easily out to be and non-representative in casein ofcase extended cloud cloud incover in the respective month. are obviously other influencing thatlight can cover the respective month. There areThere obviously other influencing parameters parameters that can impair impair light identification suchfactors as obscuring factors like smoke or fog reflections and misleading reflections identification such as obscuring like smoke or fog and misleading from snow cover, from snow cover, lightning or the aurora. cloud cover clearly considered theinmost lightning or the aurora. However, cloud coverHowever, is clearly considered theismost relevant parameter the relevantofparameter in the context the compositing process, particularly equatorial regions such context the compositing process,of particularly in equatorial regions such asin the study area of Ecuador. as the study we area of Ecuador. For recent our study we readily-processed evaluated the 6 monthly most recent available For our study evaluated the 6 most available composites (at readily-processed (at the of writing) January–June 2015 (other the time of writing)monthly coveringcomposites January–June 2015time (other monthly covering composites were only available as monthly composites were having only available as preliminary versions havingarea lower the preliminary beta versions lower quality). For thebeta Cuenca City study thequality). monthlyFor VIIRS Cuenca City areaJune the 2015 monthly VIIRS composites of May and of June 2015 feature the highest composites ofstudy May and feature the highest average number cloud-free observations (see average of cloud-free observations Table 4), number thus providing best data reliability. (see Table 4), thus providing best data reliability. Figure 99illustrates illustratesthe thelight lightintensity intensityvalues values 6 analyzed monthly composites as well as Figure forfor thethe 6 analyzed monthly composites as well as the the corresponding number of cloud-free observations at a pixel-by-pixel For theintensity light intensity corresponding number of cloud-free observations at a pixel-by-pixel basis. basis. For the light grids, grids, darker blueindicate tones indicate intensity. the cloud-free observation grids, dark blue darker blue tones higher higher intensity. For theFor cloud-free observation grids, dark blue would would represent best situation cloud cover any day month) whereas green, represent the best the situation (no cloud(no cover at any dayatduring the during month)the whereas green, yellow and yellow andindicate red colors indicatedata decreasing data reliability to fewerobservations). cloud-free observations). red colors decreasing reliability (due to fewer(due cloud-free

10


11 of 19 11 of 19

Table 4. Average number of cloud-free observations in VIIRS 2015 2015 monthly composites for the Table 4. Average number of cloud-free observations in VIIRS monthly composites forCuenca the City studyCity areastudy (greyarea indicating eventually selected selected monthlymonthly composite). Cuenca (grey indicating eventually composite). MonthMonth January January February February March March April April May June May June

Number Observation Band (DNB) NumberofofCloud-Free Cloud-Free Observation Day/Night Day/Night BandValue (DNB) Value Min Max Average Min Max Min Max Average Min Max 2 9 5.47 3.8 86.9 2 9 5.47 3.8 86.9 6 4.52 3.52 129.1 22 6 4.52 3.52 129.1 00 4 1.98 0.0 92.3 4 1.98 0.0 92.3 33 8 5.37 3.9 55.8 8 5.37 3.9 55.8 6 11 8.97 3.5 55.7 11 11 8.97 3.5 55.7 66 8.81 3.3 73.8 6 11 8.81 3.3 73.8

Theoretically, justjust a couple of high-quality observations can be can sufficient to produce appropriate Theoretically, a couple of high-quality observations be sufficient to an produce an composite product. While the monthly composites of May and June have the highest appropriate composite product. While the monthly composites of May and June have thenumber highest of cloud-free other months’other composites thus feature value number observations, of cloud-free observations, months’can composites can very thus similar feature light very intensity similar light distributions (as itdistributions is the case for(as example thefor April composite). therefore explicitly to data intensity value it is thefor case example for theWe April composite). We refer therefore reliability as an indicator as opposed to general data quality. explicitly refer to data reliability as an indicator as opposed to general data quality.

Figure VIIRS dataofofthe thefirst firstsix sixmonths months of of 2015 2015 for for the ofof average Figure 9. 9. VIIRS data the Cuenca CuencaCity Citystudy studyarea. area.Grids Grids average light intensity (top). Grids showing the number of cloud-free observations at the cell level used to to light intensity (top). Grids showing the number of cloud-free observations at the cell level used produce the average light intensity composites (bottom). produce the average light intensity composites (bottom).

11

Remote Sens. Sens. 2016, 2016, 8, 8, 114 114 Remote

12of of19 19 12

Although the May composite has the highest number of cloud-free observations on average, the Mayidentical composite highest number(see of cloud-free average, that that Although value is almost tohas thethe June composite Table 4). observations In that case,onan additional value is almost identical to the June composite (see Table 4). In that case, an additional parameter parameter should be identified to justify selection. On the one hand visual inspection of the cell-level should be identified to justify selection. could On thegive one an hand visual inspection of the cell-level distribution distribution of cloud-free observations extra indication on potential data quality. If, for of cloud-free observations could give an indication oncenter potential datanon-residential quality. If, foractivity example, example, more cloud-free observations areextra found in the city (where is more cloud-free observations are found in the city center (where non-residential activity is expected), expected), that could be beneficial given the context of the presented study. Another parameter that be beneficial contextrange, of the presented study.ofAnother the couldcould be the detected given light the intensity with detection higher parameter intensitiescould (i.e., be likely detected light intensity range, with detection of higher intensities (i.e., likely non-obscured) being non-obscured) being potentially favorable. Following the latter criterion, higher intensity levels are potentially favorable. criterion, higher are identified the June identified in the June Following compositethe as latter compared to the May intensity data set levels (see Table 4). Other in secondary composite as compared theinto Mayaccount data setinfluencing (see Table 4). Other secondary selection could take selection criteria could to take parameters that impair lightcriteria identification (as into account influencing parameters that impair light identification (as mentioned above). Data on mentioned above). Data on those parameters is usually not publicly available though. As intra-urban those parameters is usually not publicly available distributed though. As for intra-urban observationsthe at cloud-free observations at cell-level are similarly the May cloud-free and June composites, cell-level are similarly distributed for was the May and June thefactor higher intensity higher detected light intensity range eventually thecomposites, determining in detected selectinglight the June data range eventually set forwas the test study. the determining factor in selecting the June data set for the test study. To the the differences in spatial patternspatterns betweenbetween the 6 available composites, To further furtherhighlight highlight differences in spatial the 6monthly available monthly cell-by-cell light intensity deviations of every grid to the eventually selected June composite are composites, cell-by-cell light intensity deviations of every grid to the eventually selected June computed illustrated in as Figure 10. In line with the above, the May composite compositeas are computed illustrated in Figure 10.observations In line withdescribed the observations described above, matches June dataset most also inmost that closely regard.also Besides, overall patterns those the May the composite matches theclosely June dataset in thatthe regard. Besides, theofoverall cell-by-cell deviations align adequately with the corresponding grids showing the number of cloud-free patterns of those cell-by-cell deviations align adequately with the corresponding grids showing the observations (see Figure 9). The February and composites, for example, showfor the largest number of cloud-free observations (see Figure 9).March The February and March composites, example, deviations to the June grid ontoa the cell-by-cell basis, assumingly confirming the poorer reliabilitythe of show the largest deviations June grid on athus cell-by-cell basis, thus assumingly confirming those when referring the low number of cloud-free observations. March grid poorergrids reliability of thosetogrids when referring to the low number ofSpecifically cloud-freethe observations. can be considered unusable, while may be additional reasons thebe extreme lightreasons intensities Specifically the March grid can be there considered unusable, while therefor may additional for observed during the month of February (e.g., night parades and other associated carnival celebration the extreme light intensities observed during the month of February (e.g., night parades and other activities the middle of the month). associatedincarnival celebration activities in the middle of the month).

Figure 10. Cell-by-cell deviations to the selected June grid.

Table 55illustrates illustratesaaset setofof tested VIIRS threshold values using the May as alternative in Table tested VIIRS threshold values using the May data data as alternative in order order to demonstrate potential model output variation in case a different monthly composite was to demonstrate potential model output variation in case a different monthly composite was selected. selected. When the applying the 53rdthreshold percentile thresholdin(highlighted in 5)green in Table the 5) best that When applying 53rd percentile (highlighted green in Table that delivered delivered the best match to the cadastral data in the June composite, a 65:35 building use match to the cadastral data in the June composite, a 65:35 building use distribution split was obtained distribution split was obtained for the May composite, thus significantly overestimating the for the May composite, thus significantly overestimating the non-residential share. In the May data the non-residential In as thefitting Maythreshold data the(highlighted 64th percentile as approximating fitting threshold 64th percentile isshare. identified in greyisinidentified Table 5) best the (highlighted in grey in Table 5) best approximating the aggregated cadastral grid. aggregated cadastral grid.

12


13 of 19

Table 5. VIIRS distribution thresholds (original 15 arc-sec grid) and corresponding building use distribution ratios using the monthly composite for May ratios (grey indicating selected best-matching threshold, green indicating previously identified June threshold for comparison). VIIRS Distribution

ID 1 2 3 6 9

Built-up Area Distribution

Min

Threshold

Max

Percentile

Residential Use

Mixed Use

3.5 3.5 3.5 3.5 3.5

16.0 29.5 31.1 37.0 50.6

55.6 55.6 55.6 55.6 55.6

24% 50% 53% 64% 90%

32.00% 62.00% 65.00% 75.00% 96.00%

68.00% 37.00% 35.00% 25.00% 4.00%

5.2. Comparative Analysis of ISA- and VIIRS-Based Results of the Binary Land Use Classification The second aspect to be discussed is a comparison of the model output when using ISA and VIIRS-DNB data respectively. This is relevant in several aspects, most specifically (1) in terms of evaluating feasibility of continued applicability of the presented approach with the DMSP program fading out as well as (2) to assess the impact and examine expected multisided improvements due to VIIRS’ improved spatial and radiometric resolution as compared to OLS. To factor out potential influences caused by the higher spatial resolution we first compare the findings of the ISA-based analysis to those using a correspondingly aggregated 30 arc-sec VIIRS grid. Results prove to be similar in fact, with a 55th percentile threshold identified as best fit to distinguish residential and mixed occupancy areas in the VIIRS data as compared to the 50% threshold in the ISA data. In case of applying the same 50% threshold to the aggregated VIIRS composite, the obtained occupancy distribution in the correspondingly aggregated cadastral data would show a 70:30 residential-mixed split as compared to the targeted 75:25 ratio. When evaluating the degree of spatial overlap as a measure of model output accuracy, applying the respectively identified best-fitting thresholds to both data sets results in a slightly better capturing of non-residential built-up area in the binary mixed use mask that is derived from the ISA data (83%) as compared to the aggregated VIIRS data based mask (79%). If again the 50% threshold was applied to the VIIRS composite instead of the identified 55th percentile threshold, approximately 84% of the non-residential built-up area would be captured. While thereby a marginally better result is achieved in terms of capturing non-residential built-up area, the residential-mixed distribution ratio would be skewed and mixed use areas would actually be overrepresented spatially. Applying the VIIRS data in its original spatial resolution (15 arc-sec), the best-fitting threshold value to approximate the targeted 75:25 residential-mixed distribution pattern is identified at the 53rd percentile. This is slightly below the threshold value identified for the aggregated VIIRS composite (55th percentile). 76% of the total non-residential building stock of Cuenca City (3.27 of 4.3 km2 ) is captured within the derived mixed use mask. That value is below both the 79% value when using the aggregated VIIRS data and the 83% value when using the ISA data. In case of applying the initial 50% threshold for the binary classification, 79% of the non-residential building stock would be captured at a 71:29 occupancy type ratio distribution. Checking those numbers it therefore appears that using ISA data renders a better model performance than using VIIRS data both in original and aggregated form, inasmuch as more non-residential built-up area is detected in the binary masks that were derived using optimized thresholding to match residential-mixed occupancy distribution ratios. However, while a higher percentage of the non-residential built-up area is captured, ISA-derived mixed use areas are slightly more scattered. Taking VIIRS as input data source clusters the detection more in a sense that the average cell-level non-residential built-up density is higher in those binary occupancy type mask derivatives. Using the original-resolution composite, 76% of the total non-residential building stock (3.27 km2 ) is captured within 25.75 km2 , thus featuring an average non-residential built-up density


14 of 19


14 of 19

km2 .

of 12.7% per When using ISA data, 83% of the total non-residential building stock (3.57 km2 ) is With the threshold values and associated are2 . obviously rather similar for the captured within 32 km2 , thus an average density ofparameters 11.1% per km VIIRSand another interesting evaluative perspective is to derive With theISA-based threshold approaches, values and associated parameters are obviously rather similar for thea corresponding binary mask from the aggregated cadastral data and then check spatial pattern VIIRS- and ISA-based approaches, another interesting evaluative perspective is to derive a concurrence to the nightlights products. Figure 11 shows the binary classification of the aggregated corresponding binary mask from the aggregated cadastral data and then check spatial pattern cadastral data both for products. the 15 arc-sec (left) and thethe 30binary arc-secclassification (right) aggregate. The binary concurrence to (top), the nightlights Figure 11 shows of the aggregated mask separates cells that contribute strongly to the total non-residential area from cells thatmask only cadastral data (top), both for the 15 arc-sec (left) and the 30 arc-sec (right) aggregate. The binary have a marginal share. This approach is congruent to the nightlights approach in aa separates cells that contribute strongly to the total non-residential area thresholding from cells that only have sense that it aims at separating high-intensity from low-intensity cells (referring marginal share. This approach is congruent to the nightlights thresholding approach in a sense that to it “non-residential” observed parameter). The thresholds are determined in a way that, for the aims at separating as high-intensity from low-intensity cells (referring to “non-residential” as as observed nightlights data thresholding, 75:25 residential-mixed occupancy type reference ratio split is parameter). The thresholds arethe determined in a way that, as for the nightlights data thresholding, matched best-possible. For the 15 arc-sec grid the threshold is identified at 0.25% (i.e., cells the 75:25 residential-mixed occupancy type reference ratio split is matched best-possible.that Forhave the a percent-contribution to the total non-residential area of less or equal than 0.25%), while for 30 15 arc-sec grid the threshold is identified at 0.25% (i.e., cells that have a percent-contribution to thethe total arc-sec grid thearea derived threshold is 1%. Interestingly, thearc-sec difference exactly reflects non-residential of less or equalvalue than 0.25%), while for the 30 grid in thefact derived threshold the scale difference between the two datasets (i.e., factor of 4). The binary 15 arc-sec classification value is 1%. Interestingly, the difference in fact exactly reflects the scale difference between the two results in(i.e., a non-residential mask (in 15 dark blue)classification that captures 88.3%in of the total non-residential datasets factor of 4). The binary arc-sec results a non-residential mask (in 2, thus an average density of 14.9% per km2 building area of Cuenca City on an area of 25.5 km dark blue) that captures 88.3% of the total non-residential building area of Cuenca City on an area of 2 average density in the VIIRS-derived 15 arc-sec binary mask). The 30 (compared to the 12.7%/km 25.5 km2 , thus an average density of 14.9% per km2 (compared to the 12.7%/km2 average density in arc-sec mask on the hand captures 86%. the VIIRS-derived 15other arc-sec binary mask). The 30 arc-sec mask on the other hand captures 86%.

Figure 11. 11. Binary Binary classification classification of of the the non-residential non-residential cadastral cadastral built-up built-up area area aggregated aggregated to to 15 15 arc-sec arc-sec Figure (top-left) and and30 arc-sec 30 arc-sec (top-right) on the percent-contribution to the total (top-left) grids grids (top-right) based onbased the percent-contribution to the total non-residential non-residential areaBest-matching of Cuenca City. Best-matching binary classifications as derived from 15 arc-sec area of Cuenca City. binary classifications as derived from 15 arc-sec VIIRS (bottom-left) VIIRS and 30 arc-secdata. ISA (bottom-right) data. and 30 (bottom-left) arc-sec ISA (bottom-right)

For comparative comparative purposes, purposes, the the bottom bottom two two illustrations illustrations in in Figure Figure 11 11 show show the the above-presented above-presented For best-matching binary masks derived from the 15 arc-sec VIIRS and the 30 arc-sec ISA data. Visually Visually best-matching binary masks derived from the 15 arc-sec VIIRS and the 30 arc-sec ISA data. evaluatingspatial spatialdistribution distribution extent the non-residential in maps the two maps reveals evaluating andand extent of theofnon-residential class in class the two reveals interesting interesting patterns. The VIIRS-derived mask covers the south-western corner of the corresponding patterns. The VIIRS-derived mask covers the south-western corner of the corresponding cadaster-based cadaster-based non-residential mask well and misses out on the north-eastern corner whereas it is the other way around with the ISA-derived mask. VIIRS in that context seems not to detect 14

Remote Remote Sens. Sens. 2016, 2016, 8, 8, 114 114

15 15of of19 19

above-average light intensities from the Cuenca City Airport (Aeropuerto Mariscal La Mar), non-residential mask well and misses out on the north-eastern corner whereas it is the other way around whereas it is a major contributing factor in the ISA data. The latter could be explained with the with the ISA-derived mask. VIIRS in that context seems not to detect above-average light intensities inherent data configuration of ISA, which per se is more correlated with built-up area rather than from the Cuenca City Airport (Aeropuerto Mariscal La Mar), whereas it is a major contributing factor pure light intensity. in the ISA data. The latter could be explained with the inherent data configuration of ISA, which per se more correlated built-up rather than pure light intensity. 5.3.isEvaluating Model with Sensitivity via area Application of Different Spatial Urban Delineation 5.3. Evaluating Model Sensitivity via Application of Different Delineation approach with a For further evaluation of the model sensitivity we Spatial re-runUrban the implemented geospatially shrunk urban mask. While in the above-outlined implementations all the ISA and VIIRS For further evaluation of the model sensitivity we re-run the implemented approach with a grid cells were considered that fall within a pre-defined urban area of Cuenca City, now a more geospatially shrunk urban mask. While in the above-outlined implementations all the ISA and VIIRS central part of the urban agglomeration is selected. Two tests are carried out in that context. For the grid cells were considered that fall within a pre-defined urban area of Cuenca City, now a more central first one, we keep the same built-up area occupancy type distributions (75% residential vs. 25% part of the urban agglomeration is selected. Two tests are carried out in that context. For the first mixed use for VRIIS original resolution and 74% residential vs. 26% mixed use for the aggregated one, we keep the same built-up area occupancy type distributions (75% residential vs. 25% mixed use grid). In the second test, we keep the same threshold values identified above as the best match, for VRIIS original resolution and 74% residential vs. 26% mixed use for the aggregated grid). In the respectively, for each dataset (50th percentile for the ISA data and 53rd percentile for the VIIRS second test, we keep the same threshold values identified above as the best match, respectively, for data). each dataset (50th percentile for the ISA data and 53rd percentile for the VIIRS data). For the first test, the derived best-matching thresholds are now higher for both datasets. For the For the first test, the derived best-matching thresholds are now higher for both datasets. For ISA data the 55th percentile and for the VIIRS-original and aggregated grids the 63rd and 65th the ISA data the 55th percentile and for the VIIRS-original and aggregated grids the 63rd and 65th percentile are identified respectively. This was expected as predominantly residential areas in the percentile are identified respectively. This was expected as predominantly residential areas in the periphery of the city are now not included in the newly-defined urban mask and those cells periphery of the city are now not included in the newly-defined urban mask and those cells (featuring (featuring lower ISA and light intensity values) are thus missing in the histograms. The threshold lower ISA and light intensity values) are thus missing in the histograms. The threshold increment increment is higher for the VIIRS data application (roughly 10%–12%-increase) as compared to the is higher for the VIIRS data application (roughly 10%–12%-increase) as compared to the ISA data ISA data application (5%-increase). This aspect can be associated with different sensitivity of the application (5%-increase). This aspect can be associated with different sensitivity of the identified identified VIIRS and ISA thresholds due to varying histogram distributions (see Figure 12). Given VIIRS and ISA thresholds due to varying histogram distributions (see Figure 12). Given the purpose the purpose of the presented modeling, a more even histogram distribution could imply less of the presented modeling, a more even histogram distribution could imply less sensitivity in the sensitivity in the threshold determination. threshold determination.

Figure study. Figure 12. 12. VIIRS-DNB (aggregated) and ISA histograms for the Cuenca city case study.

In fact, using the ISA composite, it only takes an increment of 24% (raising the threshold from In fact, using the ISA composite, it only takes an increment of 24% (raising the threshold from 50% to 74%) to change the building occupancy type distribution ratio from 74:26 to 96:4. Regarding 50% to 74%) to change the building occupancy type distribution ratio from 74:26 to 96:4. Regarding the aggregated DNB-VIIRS it would require a 35% increment (raising the threshold from 55% to the aggregated DNB-VIIRS it would require a 35% increment (raising the threshold from 55% to 90%) to achieve the same theoretic change of the built-up area distribution. Small threshold shifts 90%) to achieve the same theoretic change of the built-up area distribution. Small threshold shifts therefore have a bigger impact when using ISA as compared to VIIRS. To illustrate and emphasize therefore have a bigger impact when using ISA as compared to VIIRS. To illustrate and emphasize this this statistically, we use a sample of 10 value pairs each for ISA, original-resolution VIIRS, and statistically, we use a sample of 10 value pairs each for ISA, original-resolution VIIRS, and aggregated aggregated VIIRS as compared to the cadastral building use distribution, and run a linear VIIRS as compared to the cadastral building use distribution, and run a linear regression (see Figure 13). regression (see Figure 13). Considering all the value pairs, in fact the original-resolution VIIRS data Considering all the value pairs, in fact the original-resolution VIIRS data shows the steepest slope in shows the steepest slope in the regression (1.2468) whereas the aggregated VIIRS data indeed show the regression (1.2468) whereas the aggregated VIIRS data indeed show the flattest slope (1.0341) with the flattest slope (1.0341) with ISA in between (1.1613). The aggregated VIIRS would thus be the ISA in between (1.1613). The aggregated VIIRS would thus be the least sensitive to threshold shifting least sensitive to threshold shifting in a sense that the building use distribution ratios would in a sense that the building use distribution ratios would accordingly deviate less from the target value accordingly deviate less from the target value (see dashed line in Figure 13). While steepest when (see dashed line in Figure 13). While steepest when considering all value pairs, the slope of the original considering all value pairs, the slope of the original VIIRS graph matches ISA almost identically 15


16 of 19


16 of 19

VIIRS graph matches ISA almost identically around the relevant target value (dashed line). Threshold around target valuewould (dashed line). Threshold shiftsimpact in theon nightlights products would shifts in the the relevant nightlights products therefore have a similar the resulting building use therefore have a similar impact on the resulting building use distribution ratios. distribution ratios.

Figure 13. 13. Plot Plotof ofresidential residentialbuilding buildinguse use ratio data percentile from histogram distribution for ratio vs.vs. data percentile from histogram distribution for ISA ISA (extended sample of Table 1), original-resolution VIIRS (extended sample of Table 2), (extended datadata sample of Table 1), original-resolution VIIRS (extended datadata sample of Table 2), and and aggregated (extended data sample 3). The lineresidential shows thebuilding residential aggregated VIIRSVIIRS (extended data sample of Table of 3). Table The dashed linedashed shows the use building use ratio for(75%) Cuenca City (75%) deriveddata. fromRegression cadastral equations data. Regression equations are ratio for Cuenca City as derived fromas cadastral are colored according colored according to the respective graphs. to the respective graphs.

In the thesecond secondtest testusing using best-matching thresholds identified the initial In thethe best-matching thresholds identified withwith the initial urbanurban mask mask (50th (50th percentile for ISA and 53rd and 55th percentile respectively for VIIRS) the newly obtained percentile for ISA and 53rd and 55th percentile respectively for VIIRS) the newly obtained built-up area built-up area occupancy type distribution for the ISA data now corresponds to a 50% residential occupancy type distribution for the ISA data now corresponds to a 50% residential and 50% mixed and use 50% mixed usetheshare fordistribution the VIIRS now data shows the distribution shows a pattern 56% share while for VIIRSwhile data the a pattern of now 56% residential and 44%ofmixed residential and the 44%original mixed resolution use considering the original resolution andthe a aggregated 48:52 ratio 30 taking in use considering and a 48:52 ratio taking in account arc-sec account thenewly aggregated arc-secarea grid. These newly derived built-up area type grid. These derived30built-up occupancy type distribution patterns are occupancy similar for both distribution patterns are similar for both data sources (ISA and VIIRS) and clearly overestimate the data sources (ISA and VIIRS) and clearly overestimate the share of mixed use area. This, again, was share of mixed use area. This, again, was expected in the same way than the first test result expected in the same way than the first test result inasmuch as in the selected central part of the urban inasmuch as ina the selected centralof part of the urban area there are a decreased number ofperiphery. residential area there are decreased number residential buildings as compared to the sub-urban buildings as compared to the sub-urban periphery. Both tests are correlated in a sense that they give indication on higher light intensity values being Both in tests are correlated a sense that they indication on higher light intensity clustered central core urban in areas of Cuenca Citygive whereas sub-urban areas feature dimmervalues lights beingconsequently clustered in central coreISA urban areason of average Cuenca as City whereas sub-urban areas feature dimmer (and also lower values) a result of higher residential densities. This lights (and consequently alsothe lower ISA values) on average aspre-identification a result of higher of residential exercise basically highlights importance of correct spatial the urbandensities. area for This exercise basically analysis. highlights the urban importance correct overspatialorpre-identification the urban subsequent intra-urban If the mask isofspatially under-defined, theofappropriate area for subsequent intra-urban analysis. If the urban mask is spatially overor under-defined, the nightlights threshold values would de- or increase respectively. appropriate nightlights threshold values would de- or increase respectively.

16


17 of 19

6. Conclusions and Outlook The presented result of the ISA data application is very interesting as it in fact backs up the prior non-evaluated assumption implemented in the Central American CDRP model to use ISA median values as a threshold for the binary land use classification of residential and mixed use areas. At the continental scale, without ground reference data as are available for the presented Cuenca City test case study, the use of the median value seemed most appropriate as it introduces the least possible subjectivity and merely separates a certain data set in high and low according to its histogram without additionally induced statistical skew. With that initially assumed median value (50%) threshold for the binary ISA classification confirmed through comparative in situ data analysis for an accurately defined urban agglomeration, the presented case study is considered very beneficial for the overall implementation process of the CDRP initiative. Also, the second re-run of the model with a geospatially shrunken, more central urban mask that showcased the correspondingly expected threshold upward shifts provides another back-up for the model validity as well as underlining the importance of accurate urban delineation in the first place. It has to be noted that with the presented Cuenca City test case, those findings have to date just been evaluated for that one particular city and caution is advised when it comes to directly transferring those conclusions to other cities. With the CDRP exposure and subsequent risk and loss models already implemented for all of Central America, further test studies can be carried out to increase the sample size of the model evaluation and also test the approach in different regional settings. Cuenca City is considered a rather typical Latin American city with regular patterns of clustered land use within the urban agglomeration. Though, in Central America, basically no major deviations are expected with regard to model applicability, it will be interesting to see testing results when extending to the Caribbean and across as well as to cities of much larger spatial urban extent. Analysis of areas further from the equator may furthermore be influenced by varying seasonal day duration as well as different cloud cover patterns, two parameters which directly affect the nighttime lights compositing. Testing VIIRS-DNB data as alternative to the DMSP-OLS-based ISA data is considered a crucial step towards a continued applicability of the model. With the DMSP program fading out, VIIRS-DNB is considered the natural successor to the OLS-based nightlights products. With certain visual improvements expected due to VIIRS’s higher spatial and radiometric resolution as compared to OLS, it is still highly valuable to get a clear idea about how these improvements eventually transfer to the binary land use classification output. One specific finding in that context refers to the stronger clustering and thus higher non-residential built-up density in VIIRS-derived binary classification as compared to the ISA-based approach. A major and often-stated benefit of VIIRS with regard to intra-urban pattern analysis is the much-improved radiometric resolution which eliminates the restricting light intensity saturation issues in urban centers in OLS data [21]. For the purpose of the presented study, however, this is not of major relevance as the OLS-derived ISA data refer to a specific radiance-calibrated nightlights product where such saturation issues have already been addressed [32]. Only two ISA datasets are publicly available, however, for the years 2000 and 2010, thus limiting potential direct applicability of the proposed approach in continuous time series analyses. Anyway, even using the annually produced and publicly available OLS stable lights-product would likely not result in a major deterioration of the binary classification as the high intensity values would still be correctly identified irrespective of their relative lower displacement in the histogram due to the saturation issue. For future studies as well as potential longer-term time series analyses, the finding is very positive in indicating that the DMSP-OLS-based ISA data and the more recent VIIRS data seem to be applicable in very similar fashion as input data sources for the residential-mixed identification model. Two main differences found using VIIRS data concern the varying threshold sensitivity and the amount of built-up detected per square kilometer of land use. The procedure of binary land use classification


18 of 19

using VIIRS is considered more flexible than using ISA, and has the potential to give a finer-scale classification of residential and mixed used in urban areas. Acknowledgments: The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of its Executive Directors or the governments they represent. Support from the World Bank’s Probabilistic Risk Assessment Program (CAPRA) and the associated Country Disaster Risk Profiles (CDRP) initiative is acknowledged. An earlier and less comprehensive version of this article showing preliminary results was presented as contribution to the 1st International Electronic Conference on Remote Sensing (ECRS-1). The concepts discussed are grounded within recent work on setting up a global-scale exposure model within the framework of the CDRP initiative, implemented in Central America and the Caribbean. The authors appreciate initial conceptual discussions within the CDRP team, specifically with Rashmin Gunasekera and Oscar Ishizawa, as well as comments from four anonymous reviewers that helped improve the quality of the paper. Author Contributions: Christoph Aubrecht developed the conceptual framework for the presented top-down method and tested geospatial implementation in various Central American countries in the context of the CDRP project initiative. José Antonio León Torres then implemented the developed approach for the Cuenca City (Ecuador) case study and performed the comparative analysis with the bottom-up cadastral data. Following up on the successful presentation of first results at the ECRS-1 conference, further research was carried out regarding the additional integration of alternative nighttime EO data sources, leading to the multi-sensor approach illustrated in this paper. Christoph Aubrecht developed that idea conceptually and José Antonio León Torres again implemented for the Cuenca City test site. The paper was initiated by Christoph Aubrecht, with major contributions from both authors in all sections. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4.

5.

6. 7. 8.

9.

10. 11.

12.

United Nations. World Urbanization Prospects: The 2014 Revision; United Nations: New York, NY, USA, 2014. World Bank. World Development Report 2014: Risk and Opportunity—Managing Risk for Development; World Bank: Washington, DC, USA, 2014. World Bank & International Monetary Fund. Global Monitoring Report 2015/2016: Development Goals in an Era of Demographic Change; World Bank: Washington, DC, USA, 2016. Gunasekera, R.; Ishizawa, O.; Aubrecht, C.; Murray, S.; Blankespoor, B.; Pomonis, A.; Daniell, J. Developing an Adaptive Exposure Model to Support the Generation of Country Disaster Risk Profiles. Earth-Sci. Rev. 2015, 150, 594–608. [CrossRef] Aubrecht, C.; Gunasekera, R.; Ishizawa, O. Urban characterization aspects and associated spatial scale considerations in the context of global-level exposure modeling: Comparative data evaluation. In Proceedings of the Second European Conference on Earthquake Engineering and Seismology (2ECEES), Istanbul, Turkey, 24–29 August 2014. Jaiswal, K.; Wald, D.; Porter, K. A Global Building Inventory for Earthquake Loss Estimation and Risk Management. Earthq. Spectra 2010, 26, 731–748. [CrossRef] Elvidge, C.D.; Tuttle, B.T.; Sutton, P.C.; Baugh, K.E.; Howard, A.T.; Milesi, C.; Bhaduri, B.; Nemani, R. Global Distribution and Density of Constructed Impervious Surfaces. Sensors 2007, 7, 1962–1979. [CrossRef] Miller, S.D.; Mills, S.P.; Elvidge, C.D.; Lindsey, D.T.; Lee, T.F.; Hawkins, J.D. Suomi satellite brings to light a unique frontier of nighttime environmental sensing capabilities. Proc. Natl. Acad. Sci. USA 2012, 109, 15706–15711. [CrossRef] [PubMed] Aubrecht, C.; León Torres, J.A. Top-down identification of mixed vs. residential use in urban areas: Evaluation of remotely sensed nighttime lights for a case study in Cuenca City, Ecuador. In Proceedings of the 1st International Electronic Conference on Remote Sensing (ECRS-1), 22 June–5 July 2015; 2015; Volume 1, p. d005. [CrossRef] Welch, R. Monitoring urban population and energy utilization patterns from satellite data. Remote Sens. Environ. 1980, 9, 1–9. [CrossRef] Elvidge, C.D.; Baugh, K.E.; Kihn, E.A.; Kroehl, H.W.; Davis, E.R.; Davis, C.W. Relation between satellite observed visible-near infrared emissions, population, economic activity and electric power consumption. Int. J. Remote Sens. 1997, 18, 1373–1379. [CrossRef] Sutton, P.; Roberts, D.; Elvidge, C.; Baugh, K. Census from Heaven: An estimate of the global human population using night-time satellite imagery. Int. J. Remote Sens. 2001, 22, 3061–3076. [CrossRef]


13. 14. 15.

16. 17. 18.

19.

20. 21. 22. 23. 24. 25.

26.

27. 28. 29. 30. 31.

32.

19 of 19

Doll, C.N.H.; Muller, J.-P.; Morley, J.G. Mapping regional economic activity from night-time light satellite imagery. Ecol. Econ. 2006, 57, 75–92. [CrossRef] Sutton, P.C.; Cova, T.J.; Elvidge, C.D. Mapping “Exurbia” in the conterminous United States using nighttime satellite imagery. Geocarto Int. 2006, 21, 39–45. [CrossRef] Elvidge, C.; Sutton, P.C.; Tuttle, B.T.; Ghosh, T.; Baugh, K.E. Global urban mapping based on nighttime lights. In Global Mapping of Human Settlement; Gamba, P., Herold, M., Eds.; CRC Press: Boca Raton, FL, USA, 2009; pp. 129–144. Zhou, Y.; Smith, S.J.; Elvidge, C.D.; Zhao, K.; Thomson, A.; Imhoff, M. A cluster-based method to map urban area from DMSP/OLS nightlights. Remote Sens. Environ. 2014, 147, 173–185. [CrossRef] Rich, C.; Longcore, T. Ecological Consequences of Artificial Night Lighting, 1st ed.; Island Press: Washington, DC, USA, 2006. Aubrecht, C.; Elvidge, C.D.; Longcore, T.; Rich, C.; Safran, J.; Strong, A.E.; Eakin, C.M.; Baugh, K.E.; Tuttle, B.T.; Howard, A.T.; et al. A global inventory of coral reef stressors based on satellite observed nighttime lights. Geocarto Int. 2008, 23, 467–479. [CrossRef] Rodrigues, P.; Aubrecht, C.; Gil, A.; Longcore, T.; Elvidge, C. Remote sensing to map influence of light pollution on Cory’s shearwater in São Miguel Island, Azores Archipelago. Eur. J. Wildlife Res. 2012, 58, 147–155. [CrossRef] Elvidge, C.D.; Imhoff, M.L.; Baugh, K.E.; Hobson, V.R.; Nelson, I.; Safran, J.; Dietz, J.B.; Tuttle, B.T. Night-time lights of the world: 1994–1995. ISPRS J. Photogramm. Remote Sens. 2001, 56, 81–99. [CrossRef] Elvidge, C.D.; Baugh, K.E.; Zhizhin, M.; Hsu, F.-C. Why VIIRS data are superior to DMSP for mapping nighttime lights. Proc. Asia-Pac. Adv. Netw. 2013, 35, 62–69. [CrossRef] Levin, N.; Johansen, K.; Hacker, J.M.; Phinn, S. A new source for high spatial resolution night time images—The EROS-B commercial satellite. Remote Sens. Environ. 2014, 149, 1–12. [CrossRef] Elvidge, C.D.; Milesi, C.; Dietz, J.B.; Tuttle, B.T.; Sutton, P.C.; Nemani, R.; Vogelmann, J.E. U.S. constructed area approaches the size of Ohio. EOS Trans. Am. Geophys. Union 2004, 85, 233–240. [CrossRef] Potere, D.; Schneider, A.; Angel, S.; Civco, D. Mapping urban areas on a global scale: Which of the eight maps now available is more accurate? Int. J. Remote Sens. 2009, 30, 6531–6558. [CrossRef] Esch, T.; Taubenböck, H.; Roth, A.; Heldens, W.; Felbier, A.; Thiel, M.; Schmidt, M.; Müller, A.; Dech, S. TanDEM-X mission—New perspectives for the inventory and monitoring of global settlement patterns. J. Appl. Remote Sens. 2012, 6. [CrossRef] Pesaresi, M.; Huadong, G.; Blaes, X.; Ehrlich, D.; Ferri, S.; Gueguen, L.; Halkia, M.; Kauffmann, M.; Kemper, T.; Lu, L.; et al. A global human settlement layer from optical HR/VHR RS data: Concept and first results. IEEE J. Sel. Top. Appl. Earth Observ. Remote Sens. 2013, 6, 2102–2131. [CrossRef] Baugh, K.; Hsu, F.-C.; Elvidge, C.; Zhizhin, M. Night lights compositing using the VIIRS day-night band: Preliminary results. Proc. Asia-Pac. Adv. Netw. 2013, 35, 70–86. [CrossRef] Municipalidad de Cuenca (GAD). Catastro Urbana de la Ciudad de Cuenca; Departamento de Avalúos y Catastros: Cuenca, Ecuador, 2015. Foody, G.M.; Cox, D.P. Sub-pixel land cover composition estimation using a linear mixture model and fuzzy membership functions. Int. J. Remote Sens. 1994, 15, 619–631. [CrossRef] Fisher, P. The pixel: A snare and a delusion. Int. J. Remote Sens. 1997, 18, 679–685. [CrossRef] Aubrecht, C.; Gunasekera, R.; Ishizawa, O.; Pita, G. The flipside of “urban”—A novel model for built-up-adjusted rural-urban pattern identification and population reallocation. In Internal CDRP Working Paper 2015; World Bank: Washington, DC, USA, 2015. Elvidge, C.D.; Baugh, K.E.; Dietz, J.B.; Bland, T.; Sutton, P.C.; Kroehl, H.W. Radiance calibration of DMSP-OLS low-light imaging data of human settlements. Remote Sens. Environ. 1999, 68, 77–88. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).