Can We Rely on VIIRS Nightlights to Estimate the Short ... · Visible Infrared Imaging Radiometer...

Policy Research Working Paper 9052

Can We Rely on VIIRS Nightlights to Estimate the Short-Term Impacts of Natural Disasters?

Evidence from Five Southeast Asian Countries

Emmanuel SkoufiasEric Strobl

Thomas Tveit

Poverty and Equity Global Practice October 2019

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Produced by the Research Support Team

Abstract

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. 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 the Executive Directors of the World Bank or the governments they represent.

Policy Research Working Paper 9052

Visible Infrared Imaging Radiometer Suite (VIIRS) night-lights are used to model damage caused by earthquakes, floods, and typhoons in five Southeast Asian countries (Indonesia, Myanmar, the Philippines, Thailand, and Viet-nam). The data are used to examine the extent to which for each type of hazard there is a difference in nightlight inten-sity between affected and nonaffected cells based on (i) case

studies of specific disasters, and (ii) fixed effect regression models akin to the double difference method to determine any effect that the different natural hazards might have had on the nightlight value. The results show little to no signifi-cance regardless of the methodology used, most likely due to noise in the nightlight data and the fact that the tropics have only a few days per month with no cloud cover.

This paper is a product of the Poverty and Equity Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at [email protected].

Can We Rely on VIIRS Nightlights to Estimate the

Short‐Term Impacts of Natural Disasters?

Evidence from Five Southeast Asian Countries

Emmanuel Skoufias*

Eric Strobl**

Thomas Tveit**

JEL Classification: Q54, C63, R11, R5,O18 Keywords: Remotely Sensed Data, Natural Disasters, Damage Index, Floods, Earthquakes, Typhoons

*The World Bank Group, Washington, DC USA

** University of Bern, Switzerland

2

Introduction

Natural hazards and the damages caused by them lead to headlines in the news now more than ever. Part of the reason for the increased reporting is due to the focus on climate, but it is also caused by larger economic damages than in prior years. The latter point is due to the growing global population and the increasing economic value in settled places. With this value increase comes a need for updated local data on where people live and the economic activity there. This is particularly true in the wake of a devastating natural hazard, when it can be difficult to assess damages quickly and accurately.

A recent method to assess damages and economic activity is to use remotely sensed data. As a proxy for economic activity one has seen widespread use of nightlights (Henderson, Storeygard and Weil 2012) (Gillespie, et al. 2014) (Skoufias, Strobl and Tveit 2017) (Hodler and Raschky 2014) (Michalopoulos and Papaioannou 2014) (World Bank 2017). The modeling and measurement of natural hazards has also been extensively covered in the literature for different disaster types such as earthquakes (De Groeve, et al. 2008), typhoons (Emanuel 2011) and floods (Wu, Adler and Tian, et al. 2014) to name a few.

The aim of this paper is to provide a refined model of a previous study (Skoufias, Strobl and Tveit 2017) where nightlights were used as a proxy for economic activity and then combined with damage indices to find the economic impact. The novelty is that a new, more spatially and temporally detailed data set for nightlights, the Visible Infrared Imaging Radiometer Suite (VIIRS) product from the National Oceanic and Atmospheric Administration (NOAA), following the work of Elvidge et al. (2017), is employed. In addition, settlement layers have been used to remove nightlight cells with zero population to reduce the impact that “empty” cells could have on the analysis.

Five countries in Southeast Asia have been chosen as case studies due to their general lack of local level economic data and prevalence of different natural hazards: Indonesia, Myanmar, the Philippines, Thailand and Vietnam. These countries represent a large portion of the total GDP, population and area of the region. In addition, they experience earthquakes, typhoons and floods to a varying degree.

The VIIRS data have a number of advantages over the widely used, but now discontinued, Defense Meteorological Satellite Program (DMSP). Firstly, they have a higher resolution, 450m by 450m compared to 1km by 1km for the DMSP. Furthermore, the VIIRS data are released monthly compared to the annual DMSP data, making it possible to compare month‐on‐month effects of natural disasters. Finally, the DMSP data are normalized and capped at 63, whereas the VIIRS data have no upper limit. However, the VIIRS data exhibit a lot more noise than what is found in the DMSP data. Part of this is due to stray light corrections that can lead to negative light values for some months. Another issue, prevalent in both data sets, occurs due to the many days with cloud cover that happens in the tropics. As a matter of fact, several months have no days with a measurement. Importantly, both the DMSP and the VIIRS data have shown strong correlation with local GDP (World Bank 2017) (Henderson, Storeygard

3

and Weil 2012) and this paper will take advantage of this to use VIIRS as a proxy for economic activity.

The settlement layers used in the analysis are from WorldPop (WorldPop 2013) (Worldpop 2016) and CIESIN (Facebook Connectivity Lab and Center for International Earth Science Information Network ‐ CIESIN ‐ Columbia University n.d.). Both sets are high resolution layers that show where people live through a combination of satellite images and census data. There are two reasons for including the settlement layers. Firstly, the data are used as a way to clean the VIIRS data by excluding cells that are unlikely to contain economic activity and population. The use of secondary data to improve recognition of human activity has previously been done for poverty (Jean, et al. 2016) and for urban markets (Baragwanath, et al. 2019). Secondly, the population data are used to distinguish between rural and urban areas.

For the disaster modeling, we model three disaster types, earthquakes, typhoons and floods, by following closely the methodology from Skoufias, Strobl and Tveit (2017) for earthquakes and floods and Strobl (2012) for typhoons. These damage indices are then combined with the other data sets and two types of analysis are performed on this combined set. The first type is one case study for each hazard type. The choice fell on typhoon Haiyan, which hit the Philippines in 2013, the 2016 Aceh Earthquake and the 2017 floods in Southern Thailand. All three disasters caused significant local damage and were for the most part spatially and temporally separate from other disasters. The case studies consist of a simple graph showing the difference in mean nightlight values for the nightlight cells that were hit by the hazard and the cells that were not hit over a period starting 12 months before the hazard and until 12 months after. The second analysis moves the focus from specific events to a national level and hazards that occurred between June 2012 and March 2018. To find any short and medium‐term impacts that each disaster type has, a fixed effect regression with 12 monthly lags is performed.

Overall, the results are not particularly significant for either analysis. Two case studies of typhoons in the Philippines exhibited the strongest results, where one found a clear drop in mean nightlight during the month that Haiyan struck and similarly a clear drop when Rammasun struck in July 2014. The 2017 floods in Thailand also exhibited a drop, but compared to noise in other months it was rather small, whereas the 2016 Aceh Earthquake exhibited no difference between cells hit by it versus those that were not affected. Part of the reason for these results could be that an influx of aid and personnel would lead to a net effect of zero, but given the noise that is exhibited overall in the VIIRS data it would be hard to draw any more confident conclusions in this regard. The issue of noise in the underlying monthly data has been well illustrated in Hu and Yao (2019), where the monthly VIIRS data showed much lower coorelation with GDP than annual VIIRS and DMSP data.1

The fixed effects regressions were also inconclusive, as there was little if any significance across disaster types and months. Several methods were tried to correct for the noise, such as dropping any negative values, setting any value below a certain threshold to zero (both for

1 The annual VIIRS data were not used in this study, as they are currently only available for 2 years

4

absolute and real values), running them only for cells with populations above a certain threshold as well as logaritmic hyberbolic sine transformations. However, the results did not improve or change. There could be several reasons for these results. The most likely being that the VIIRS data are noisy by nature and, due to the countries being close to the equator, many areas are often covered by clouds, reducing the quality of the measurements. The VIIRS data also consist of much noise near zero, meaning that a transformation or cleaning of the data was necessary, although perhaps not the best one was used. Another possibility is that the methodology or measurements for the different hazard types are inaccurate. Finally, the choice of regression model might also cause estimation problems. Overall, different methods were employed to alleviate these concerns, but none yielded the expected results.

The remainder of the paper starts with an overview of the countries and the base data, followed by a general methodology overview, the three disaster types with results and finally a conclusion.

I. Overview, Common Data and Summary Statistics

A. Countries

We have chosen five different Southeast Asian countries: Indonesia, Myanmar, the Philippines, Thailand and Vietnam. These five countries have been chosen for several reasons. Firstly, they represent roughly 75 percent of the Southeast Asian GDP,2 providing a good representation of the region as a whole. Secondly, they constitute over 80% of the total landmass and 90% of the total population in Southeast Asia. Finally, they all represent different economies and their economies are affected differently by natural hazards.

2 According to the World Bank Group Research https://data.worldbank.org/indicator/ny.gdp.mktp.cd.

5

Figure 1: Countries in Southeast Asia

Indonesia. The largest country in terms of GDP, area and people is Indonesia. It has more than 40 percent of the total population in Southeast Asia, with more than 260 million inhabitants. It is also the country with the highest number of natural disasters. Geographically Indonesia is mostly made up of large and – in some cases – densely populated islands.

According to the Indonesian National Disaster Management Authority (BNPB), there were more than 19,000 natural hazards in the period 2001 ‐ 2015 (National Disaster Management Agency 2016), making Indonesia a useful country for any natural hazard analysis. The most frequent disasters are floods and landslides (52 percent), strong winds (21 percent) and fires (15 percent), while the most damaging ones are earthquakes, tsunamis and volcanic eruptions, which all cause major damage to buildings and infrastructure in addition to the human casualties.

The tropical climate of Indonesia often leads to annual floods. The BNPB data registered more than 10,000 incidents of floods or landslides leading to more than 3,500 fatalities from 2001

6

through 2015. During the period from 1985 to 2016, the Dartmouth Flood Observatory (DFO) registered 3,808 floods of magnitude 4 or more and 1,175 floods of magnitude 6 and up.3

Myanmar. The second country used in our analysis is Myanmar. It has the smallest GDP and population of our five countries, while at the same time being the second largest. During the period 2006‐2016, the UN Office for the Coordination of Humanitarian Affairs (OCHA) mentions several major natural hazards affecting Myanmar (OCHA 2016).

There were three large earthquakes in 2011, 2012 and 2016, affecting more than 30,000 people and killing 100. Furthermore, a major flood occurred in 2015, displacing 1.7 million people in Western Myanmar and killing 172 people. Finally, more than 2.5 million people were affected by strong cyclones in 2008, 2010 and 2013.

Philippines. The Philippines is – like Indonesia – a country consisting of thousands of islands ranging in size from tiny and deserted to large and densely populated. It is the second smallest country in our sample, with the second highest population just in excess of 100 million. It has the third largest GDP.

According to the Global Facility for Disaster Reduction and Recovery (GFDRR), the Philippines is at high risk from several types of natural hazards (GFDRR 2019). Prime among them are cyclones, where an average of 20 make landfall every year. In 2013, typhoon Yolanda led to 6,000 casualties and damaged more than 1.1 million houses. The Philippines are also exposed to earthquake and flood risks.

Thailand. As for Thailand, it is the third largest country, fourth most populous and boasts the second largest GDP of our countries. In terms of disasters, Thailand is mostly at risk to floods and typhoons, but earthquakes happen occasionally, with the largest recorded being a 6.1 magnitude earthquake that occurred in May 2014. The lower risk compared to our other countries is also mentioned by INFORM in their country profile of Thailand (INFORM 2019).

Vietnam. The final country is Vietnam, which is the smallest, but has the third highest population just below 100 million people. In terms of GDP, Vietnam rank fourth among our five countries. When it comes to disaster risk, they are on par with Thailand in that INFORM ranks them low and exposed primarily to floods and typhoons (INFORM 2019).

B. VIIRS

To find and identify areas with economic activity and their exposure to natural hazards, satellite images of nighttime lights will be used as a proxy. This is due to the highly localized nature of natural hazards and how they affect economic activity within populated areas. In an optimal scenario, one would prefer highly spatially disaggregated economic data, but lacking

3 Magnitude is defined as: M = log(D * S * AA), with D being the duration of the flood in days; S is the severity on a scale consisting of 1 (large event), 1.5 (very large event) and 2 (extreme event); and AA is the size of the affected area in square kilometers. Flood events registered by DFO have mainly been derived from news and governmental sources.

7

this, one can use nightlights. This methodology has been used in several papers already, with significant results (Henderson, Storeygard and Weil 2012) (Gillespie, et al. 2014) (Hodler and Raschky 2014) (Michalopoulos and Papaioannou 2014). In Henderson, Storeygard and Weil (2012), Indonesia is used as an example of using nightlights to capture an economic downturn following the Asian financial crisis in the late 1990s. Their results show that swings in GDP change can generally be captured. Nevertheless, one has to account for factors such as cultural differences in light usage, latitude and gas flares. In our case this is unlikely to affect our results since nightlights are used to capture exposure within a country rather than across countries.

All the previously mentioned studies use a prior iteration of nightlight images, the DMSP satellite images. The approximate cell grid size of these images is 1 by 1 kilometer at the equator and the data provided are annual.

This paper utilizes a newer nightlight data set, the VIIRS Day/Night Band (DNB) provided by The Earth Observations Group (EOG) at NOAA/NCEI. The data are produced following the methodology of Elvidge, et al. (2017). These data are provided starting in April 2012 and go till the present, making the time series much shorter than the DMSP data which start in 1992 and go through 2013. However, the VIIRS DNB images from EOG are monthly, whereas the DMSP data are annual. Furthermore, the images have 15 arcseconds grids (approximately 450 meters at the equator) compared to the 30 arcseconds of DMSP. The VIIRS coverage spans from 75N latitude to 65S around the entire globe, meaning that all of Southeast Asia is included. The VIIRS data have also seen usage in the economics and disaster literature. Firstly, Chen and Nordhaus’s (2015) analysis finds promising results for VIIRS as an economic and population indicator, also when comparing to the DMSP product. Secondly, Zhao, et al. (2018) used the underlying NPP‐VIIRS DNB Daily Data to analyze selected natural disasters. They found that the images were useful for detecting damages and power outages, but that cloud coverage was a major limitation in the assessment.

More recent research utilizing the VIIRS nightlights also points to the limitation of monthly VIIRS in the detection of GDP and urban markets. Hu and Yao (2019) show how monthly VIIRS data have very low correlation for low income countries.4 The overall correlation is significantly better for middle income countries and annual, corrected data. Even when using the annual mean of monthly data, the results are worse than for annual VIIRS and DMSP.

The data output consists of two variables: the average light radiance from DNB and the number of cloud free days. An important note regarding cloud free days is that for the tropical areas in Southeast Asia one will have months where there are no cloud free days and hence no radiance measurement. To account for months with no radiance value, this paper has interpolated between the month before and after. Furthermore, for cells with little light, one often finds negative light values due to airglow contamination (Uprety, et al. 2017). There is no established standard for how to interpret these measurements, but we did any analysis

4 If one includes all the low income countries, the correlation is negative (‐0.02), but it is heavily skewed by the poorest African nations. With

the exclusion of the Central African Republic, the correlation improves to 0.12.

8

using different methodologies such as setting negative values to 0 or using absolute values with different thresholds for counting a value as 0.

To provide some further details into the noise and distribution of the VIIRS data, Figures 2 and 3 show the distribution of VIIRS values below 25 for all countries and their populated areas and the number of cloud free days per month. Looking at Figure 2 first, one sees a clustering on both sides of 0, with the bin containing 0 and very small values constituting more than 15 percent. Furthermore, more than 13 percent of the total points have a negative value and 46 percent more have a value between 0 and 0.3, meaning that almost 60 percent of all points have negative or very small values. It is also worth noting again that these are only points that have been identified as being populated,6 meaning that they should be relatively free of disturbance from non‐human sources. Despite this, there are still very small deviations from 0.

Even when limiting the light to cells with population over 50 (approximately 30 percent of populated cells), the distribution stays similar with 85 percent of light values being below 2, 75 percent below 1, 46 percent below 0.3 and 5 percent being negative. This distribution pattern follows through to high population numbers. When looking at population numbers above 1,000 (less than 12 percent of the total) 17 percent of the points still have nightlight values below 0.3 albeit only 0.3 of points have values below 0. The question will then be if VIIRS is only useful for very urbanized areas, if at all.

5 Values below 2 constitute approximately 95 percent of the total. This was done due to the maximum VIIRS value being 17,699 and hence a

density graph would be meaningless going from ‐1.5 to 17,699. Also, any bin above 2 is very small and would not contribute to the graph.

6 See the next section for details on population layers.

9

Figure 1: VIIRS Value Distribution for VIIRS values below 2

Source: Authors’ estimates based on VIIRS and population layer data (see text for details)

Figure 3 provides an overview of the distribution of cloud free days. The first thing of note is that there are no points or months that have more than 20 cloud free days, meaning that no month has used more than 2/3 of the days to get a monthly mean value. The median is 4 days and the mean is 6, and in almost 11 percent of months there were no cloud free days. Overall, the implication is that the monthly values in reality consist of a rather small subsample of the monthly light. It might not matter much for a non‐climate related hazard such as earthquakes, but for a flood, which is highly correlated with clouds, it might mean that the mean value is based on values before or long after the event has occurred.

10

Figure 3: Days with No Cloud Cover in a Month


C. Population Layers

An important aspect of capturing economic activity through nightlights is to identify areas where there are people, hence avoiding stray lights or other sources of light that are not connected to human activity. For example, Baragwanath et al. (2019) used daily MODIS data to identify the extent of urban markets in India. To find these areas, two sources of human settlement layers are utilized.

The WorldPop data sets have been used to identify settlement areas in Myanmar and Vietnam (WorldPop 2013) (Worldpop 2016). These data sets have a spatial grid cell resolution of approximately 100 meters at the equator and estimate the number of persons per square. Estimates are provided for 2010, 2015 and 2020, but for this paper we have only used the 2015 estimates. To construct the population files, national totals have been adjusted to match UN population estimates. To do this adjustment, a Random Forest model has been used to construct a weighted population density layer, which is used as a basis to place the population as closely as possible to its real geographical distribution.

11

For Indonesia, the Philippines and Thailand high resolution settlement layers from Facebook Connectivity Lab and Center for International Earth Science Information Network ‐ CIESIN ‐ Columbia University was used. These layers are produced at a spatial resolution of 1 arcsecond (~30 meters). The underlying data are based on 2015 and combines census data with satellite images from DigitalGlobe. The population is allocated according to subdivision censuses once settlements have been identified from the satellite images.

In addition to identifying economic activity, the settlement layers will be used to identify rural and urban areas based upon population density.

II. General Empirical Strategy

The goal is to analyze the effect that natural hazards might have on the economic activity in an area. To try to tease out this effect, a data set was constructed for each natural hazard type. These sets contain localized nightlight values and damage indices for each hazard type, the total population in each light cell and administrative boundary data linking each nightlight cell to different administrative levels.

The new data sets are used in two different sets of analysis. First, a simple case study on a singular large event is performed. The case studies are done on the Aceh earthquake in 2016, the Haiyan Typhoon and finally for the 2017 floods in Thailand. All of these events had severe damages in the affected areas and one would expect that the nightlight values would change significantly following these events. As for the actual analysis, it is a graph of nightlight means of cells in affected areas, split between cells that experienced damage versus those that did not.

The second analysis is a fixed effects regression for each country and natural hazard type to see if any national effects exist. The regressions are lagged for the 12 months following the hazards to allow for any short‐ and mid‐term effects to materialize. The general methodology can be seen in Figure 4.

12

Figure 4: Flow chart for methodology

A. Constructing the Data Sets

The approach we have chosen is somewhat similar to what was done in Skoufias, Strobl and Tveit (2017), where the authors constructed damage indices, weighted with nightlights and aggregated up to annual values at a province or municipality level. However, the method this time also provides significant differences.

Firstly, the nightlight data used as a proxy for economic activity is monthly – and not yearly – providing a better picture of how nightlights are affected by natural hazards in the short term. Furthermore, the spatial resolution is much higher, with the VIIRS data providing details down to 450 meter squares at the equator compared to 1 kilometer squares for the DMSP.

13

Secondly, by using high resolution population layers to identify areas with economic activity some noise from adjacent non‐economic areas is removed. Hopefully providing a better data set to pinpoint where a change in light value would be due to a change in economic output and not due to stray light or other light sources such as forest fires. One potential downside of this is that some agricultural land will not be included in the analysis. Finally, instead of aggregating up to a provincial or municipality level, the data are kept at a cell level.

More specifically with regards to the methodology, one started with localized damage data modeled from actual intensity measures. These values were matched with any intersecting VIIRS cell and assigned to the corresponding month. For light cells that intersected several damage cells an average value was used. As earthquake damage estimates were modeled based on centroids, only the centroid intersection was used.

Once the nightlight and damage data had been merged, population data was included. The population data sets were aggregated up to the same cell size as the VIIRS data and then matched to find the total population in each nightlight cell. Due to slight differences between the HRSL and WorldPop data sets, two different cutoffs for populated cells were utilized. As HRSL specifically provide settled areas and set other areas to having 0 population, any VIIRS cell with no settled areas in it, would be excluded from the final data set. As for the WorldPop data, any VIIRS cell with less than 5 people in it was cut.

Finally, for potential future analysis, the VIIRS cells were assigned to administrative subdivisions such as provinces and municipalities. In cases where a cell crossed province borders, the province that contained the centroid was used.

B. Case Study Methodology

Once the data sets were constructed, one case study for each natural hazard type was chosen to graph the effect the hazard would have on the local nightlight values. The choice of hazards was based on reported damage and “fame”. In the case of floods, they were also picked based on how “isolated” they were, i.e. some of the largest events were lasting for months on end potentially confounding the effect of floods in different months.

To check for any effect from the disasters, two graphs were constructed for each event; one with the mean of the nightlight values of cells that were struck by the event and one with the cells that were not affected. Furthermore, the analysis focused on the affected regions – state or province level ‐ and not the national averages.

C. Regression Methodology

The second set of analysis consists of regressions which were run to determine any effect that the different natural hazards might have had on the nightlight value. The methodology chosen were that of fixed effects, correcting for time and spatial effects. To correct for potential heteroscedasticity, Driscoll‐Kraay standard errors are used (Driscoll and Kraay 1998).

14

𝐿 , 𝛽 Σ 𝛽 𝐸𝐷 , , 𝜃 𝑒 ,

where 𝐿 , is the light level in cell i in month t and 𝐸𝐷 , , represents the vulnerability curve

value in the same cell and at the same time. Lags are allowed from month t to t‐12. β is the intercept, θ are the cell fixed effects and 𝑒 , is the error term.

III. Earthquakes

To construct damage indices for earthquakes, this paper uses the same methodology as in Skoufias, Strobl and Tveit (2017), which in turn is based on modeling from contour maps generated by seismological ground stations (International and Regional Development 2001) (Agency 2006) (De Groeve, et al. 2008).

In short, these contour maps are ShakeMaps from USGS, which are automatically generated maps providing several key parameters following an earthquake, such as peak ground acceleration (PGA), peak ground velocity (PGV) and modified Mercalli intensity (MMI), are used as a base for localized impact. More specifically, the ShakeMaps use data from seismic stations that is interpolated using an algorithm which is similar to kriging. To model the intensity in a given coordinate, the model also takes into account ground conditions and the depth of earthquake. The ShakeMaps are interpolated grids with point coordinates spaced approximately 1.5 kilometers apart (0.0167 degrees).

With regards to the output data, PGA is a measure of the maximum horizontal ground acceleration as a percentage of gravity, PGV is the maximum horizontal ground speed in centimeters per second and MMI is the perceived intensity of the earthquake, a subjective measure. It is assumed that damage starts at an MMI level of V and a PGA of 3.9 percent of gravity (g). These damage levels are found for California in (Wald, et al. 1999), but the relationship has been found for other areas in the United States in Atkinson and Kaka (2006) and Atkinson and Kaka (2007) and for places such as Costa Rica (Linkimer 2007) and Japan, Southern Europe and the Western United States (Murphy and OBrien 1977). It is worth noting that the numerical relationship differs from region to region.

The different measures are largely interchangeable, and in the International and Regional Development 2001 report, they use PGA to measure damage, pointing to the fact that PGA, unlike MMI is an objective measure, implying that MMI is not easy to obtain reliably across the globe. Also, for large scale modeling, where it is unfeasible to model local conditions precisely, PGA serves as a good proxy for intensity of earthquakes.

For the actual construction of the damage index, two types of data will be used; the intensity data ‐ expressed as PGA ‐ and building inventory data. The building type data stems from the USGS building inventory for earthquake assessment, which provides estimates of the proportions (based on total number of buildings) of building types observed by country; see

15

Jaiswal and Wald (2008). The data provide the share of 99 different building types within a country separately for urban and rural areas. For Indonesia the building type information was compiled from a World Housing Encyclopedia (WHE) survey. The WHE survey uses fraction of population who lives or works in buildings of different types as their definition of how the building mass is split up. Without any other information available, we use this as an indication of the distribution of building types and mass, but, necessarily, assume that the distribution is homogenous within urban and rural areas.

Damage curves by building type are derived from the curves constructed by the Global Earthquake Safety Initiative (GESI) project; see (International and Regional Development 2001). More specifically, buildings are first divided into 9 different types across a wide range of building materials from wood and steel through to lightweight shacks. Each building type itself is then rated according to the quality of the design, the quality of construction, and the quality of materials. Total quality is measured on a scale of zero to seven, depending on the total accumulated points from all three categories. According to the type of building and the total points acquired through the quality classification, each building is then assigned one of nine damage curves which provides estimates of the percentage of building damage for a set of 28 peak ground acceleration intervals.

In order to use these damage curves, we first allocated each of the 99 building types given in the USGS building inventory to one of the 9 more aggregate categories of the GESI building classification. However, to assign a building type its quality‐specific damage curve we would further need to determine its quality in terms of design, construction, and materials, an aspect for which we unfortunately have no further information. We instead assume that building quality is homogenous across building type and construct eight different damage curves with different quality rating scenarios (ranging from 0 to 7). Initial analysis showed some effect based on the quality assumption, but lacking any meaningful information we decided to use a mean quality assumption of 4 across the countries.

To model estimated damage due to a particular earthquake event the data from the ShakeMaps and GESI are used. Then, one identifies the value of peak ground acceleration that each nightlight cell experiences by matching each earthquake point with its nearest nightlight cell. If the cell is further away than 1.5 kilometers or if it experiences shaking (PGA) of less than 0.05g the value is set to 0. In order to derive a cell i specific earthquake damage index, ED, the following equation is applied:

𝐸𝐷 , , 𝐷𝑅 , ,, 𝑞 0, … ,7

where 𝐷𝑅 is the damage ratio according to the peak ground acceleration, 𝑝𝑔𝑎, and the urban‐rural qualification 𝑘 of cell 𝑖, defined for a set of 8 different building quality 𝑞 categories.

Once the damage index has been constructed, the different data sets are combined and a final data set with the nightlights, population and earthquake data are made, ref Figure 2. With this combined data set, we do a case study of the 2016 Aceh Earthquake and we run the fixed effect regression:

16

𝐿 , 𝛽 Σ 𝛽 𝐸𝐷 , , 𝜃 𝑒 , ,

which is the same as in section II.c. In short, it gives the coefficients for the effect that the earthquakes have on the nightlights for the month when the earthquake happens and the 12 months after the earthquake. The regression is run for all countries bar Vietnam, which did not experience any earthquakes during our period.

Results. The starting point for results is the case study of the 2016 Aceh Earthquake. This is the single most damaging event in our study period and occurred in the Indonesia province of Aceh in the morning of 7 December 2016. The magnitude was measured to be 6.5Mw with number of fatalities being 104 and more than 80,000 were displaced. The affected area was rather small, primarily affecting the Pidie Jaya Regency towards the northwestern part tip of Sumatra. However, within the Regency it was stated that as much as 30% of the area was affected. This makes it a good case to study, with a highly localized event that causes severe destruction.

Figure 5: 2016 Aceh Earthquake – Mean of nightlight cell values in the province of Aceh split between cells that were affected by the earthquake and those that were not


Figure 5 consists of two graph lines, with the green one being the mean VIIRS nightlight value for cells that were not damaged by the earthquake and the blue one being the same, but for cells that were struck by the earthquake.

The first thing to note is that the values fluctuates a lot and for the month before the earthquake the values almost drops to zero. Unaware of any other disasters striking that month, the underlying VIIRS data for all of Indonesia were checked. It turns out that there was a nationwide drop from October to November followed by an increase from November

17

to December. There were no long‐lasting, nationwide power outages reported, and although there was an increase in the number of cells with no cloud free days, it cannot explain a deviation of this magnitude. The most likely explanation is some noise in the underlying VIIRS data causing the November 2016 data to deviate from other months.

Regardless of the November data, there does not seem to be a big drop or anything else that causes the mean of the cells struck by the earthquake to deviate from the mean of the non‐affected cells. This could be due to aid and relief efforts, but given the noise already seen in the data, it is more likely that VIIRS has been unable to capture the effects the Aceh Earthquake had on the local economy.

The second set of analysis is the fixed effect regression. The above regression was run with a number of different specifications. The key assumptions were related to the following areas:

1. Distinction between rural and urban areas to decide which vulnerability curve was applicable

2. Treatment of VIIRS values below 0

With regards to deciding between rural and urban areas, no national data are available to the authors’ knowledge. Furthermore, it would be preferable to use the same definition across all countries to have comparable numbers. As such, the regressions were run with numerous specifications ranging from the most extreme cases where the countries were assumed to be all rural or all urban to cases where any cell with less than 10, 50 or 100 people were assumed rural. Overall, this assumption did not affect the results much in terms of significance, but there were some minor shifts in coefficients. Due to the lack of impact from this assumption, it is in the following assumed that any VIIRS cell (450m by 450m) containing fewer than 10 people is considered rural, while any cell with more is considered urban.

As for the treatment of VIIRS values below 0 ‐ or close to 0 for that matter – there is no clear consensus in literature. The methodology for the construction of the light value is such that a dark offset is deducted from the raw day/night value. This offset is sometimes severely impacted by airglow leading to negative light values. To account for this several options were explored. Firstly, the regressions were run with any light value below 0.1 set to 0. Then absolute values were used with any absolute value below thresholds of 0.1, 0.3 and 0.5 set to 0. Once more this did not seem to affect the results a lot, although there were some minor shifts. With no possible way of cleaning the underlying data, a threshold of 0.3 absolute value was chosen as the base regression case.

The results for the different countries are shown in Figures 6 through 9. Vietnam did not yield any results as it did not experience any earthquakes of sufficient strength during our period of analysis.

18

Figure 6: Indonesia earthquake coefficients with 95% confidence intervals


Figure 6 shows the results for Indonesia. There were no significant effects for earthquakes on nightlights neither in the month that an earthquake hit nor in any of the following months. This might be due to several factors. As Indonesia is often struck by earthquakes, they might have good and efficient ways of dealing with the damage from earthquakes. It could also be that the regression specifications were not good for this analysis on a country such as Indonesia. Due to the size and geography, there might be large local differences, something which should be accounted for with a clustering option. As of now, it has not been possible to run this due to the size of the data set with such an option. Secondly, it might be that it is necessary to increase the PGA assumption regarding when an earthquake is damaging or not. It might be that there is limited damage from the small earthquakes in Indonesia due to the prevalence of these. After all, the average impact an earthquake had in a nightlight cell was approximately 6 percent, while the maximum damage was almost 80 percent, implying that there was a vast range in the local damage caused by earthquakes.

19

Figure 7: Myanmar earthquake coefficients with 95% confidence intervals


Figure 7 shows Myanmar, which experienced a fair number of earthquakes during the period. However, there were no significant coefficients here either, apart from a strange positive blip 11 months after the earthquake. This is most likely due to the small impact that each earthquake had. The maximum damage experienced in a nightlight cell was barely in excess of 5 percent.

Figure 6 shows the results for the Philippines and here there is actually a significant and positive effect on the nightlights in the month that the earthquake happens. This might be due to a good and efficient local infrastructure. However, with no further effect later on, it is not clear what might be causing this early light influx.

20

Figure 8: Philippines earthquake coefficients with 95% confidence intervals


Finally, Figure 9 shows the results for Thailand. These are highly significant for several months. However, with only 169 affected nightlight cells and a maximum damage of 8 percent, it is doubtful that these results show much about the economic impact an earthquake has on local economic output in the short‐ or medium‐term.

21

Figure 9: Thailand earthquake coefficients with 95% confidence intervals


Overall, it seems as if earthquakes do not affect local economic output in Southeast Asia as much as one would expect. In the cases of Thailand and Myanmar this might be due to the limited impact from earthquakes overall, whereas in the Philippines and Indonesia one would expect some effect. It might be that local cloud cover leads to poor measurement conditions or it could be that the regression specifications are not correct.

IV. Typhoons

To model typhoons, a damage index from Emanuel (2011) is used. It assumes that a fraction of property is lost or damaged when wind speeds surpass a threshold in a cubic manner.7 Formally our destruction index is given by:

𝑓 𝑣

1 𝑣

where 𝑓 is the fraction of property lost or damaged and 𝑣 is:

𝑣 ≡𝑚𝑎𝑥 𝑉 𝑉 , 0𝑉 𝑉

7 Damages are related to wind speed in a cubic manner due to the nature of energy dissipation of the hurricane.

22

where 𝑉 is the wind speed, 𝑉 is the wind speed at and below which no damage occurs (set at 92.6km/h) and 𝑉 is the wind speed at which half the property is destroyed (set at

203.7km/h).

The wind speed data we use follows Strobl (2012), which in turn is based on a wind field model developed by Boose, Serrano and Foster (2004), which in turn is based on Holland (1980). The base equation is given as:

𝑉 GF 𝑉 S 1 sin 𝑇 𝑉2

𝑅𝑅

𝑒𝑥𝑝 1𝑅𝑅

/

where 𝑉 is the wind speed at point 𝑃, 𝑉 is the maximum sustained wind velocity anywhere in the hurricane, 𝑇 is the clockwise angle between the forward path of the hurricane and a radial line from the hurricane center to the point of interest (the centroid of a Parish), 𝑃, 𝑉 is the forward velocity of the hurricane, 𝑅 is the radial distance from the center of the hurricane to point 𝑃, and 𝐺 is the gust wind factor (water = 1.2, land = 1.5). Finally, 𝐹 is a scaling parameter for surface friction (water = 1.0, land = 0.8), 𝑆 is the asymmetry due to the forward motion of the hurricane (1.0) and 𝐵 is the shape of the wind profile curve (1.2). These parameter values have been verified in Boose, Chamberlin and Foster, Landscape and Regional Impacts of Hurricanes in New England (2001) and Boose, Serrano and Foster (2004).

The source for the localized wind speeds is the IBTrACS database that provides the strength and tracks every 6 hours of all typhoons that affected Southeast Asia during the period.

As for earthquakes, the damage index and the different data sets are combined and a final data set with the nightlights, population and damage data are made, ref Figure 4. This is used for a case study of typhoon Haiyan in 2013 and in the fixed effect regression:

𝐿 , 𝛽 Σ 𝛽 𝑓 , 𝜃 𝑒 , ,

which is the same as in section II.c, giving typhoon effect coefficients for the event month and the 12 months following. As the Philippines and Vietnam were the only two countries to experience any damaging typhoons during our period, the regression was only run for those two countries.

Results.—Once again the first analysis will cover a case study, the typhoon Haiyan. It is also known as typhoon Yolanda on the Philippines and was active from November 3, 2013 through November 11, 2013. It made landfall in the eastern parts of the Philippines on November 7 with wind speeds that were 305km/h, a then‐record velocity. It continued across the middle islands of the Philippines leaving the islands on November 8. It is widely known as the deadliest and costliest typhoon to ever hit the Philippines in modern times, killing 6,300 people and causing damages worth US$2.2 billion. As a side‐note, in July 2014 typhoon Rammasun struck some of

23

the same areas and costing US$885 million in total damages, the third costliest typhoon ever to hit the Philippines.

Figure 10: Case study of Typhoon Haiyan and Rammasun showing the mean VIIRS value for cells that were affected by typhoon Haiyan and those that were in the same province but

that were not damaged


Figure 10 shows the lines for mean nightlight values before and after Haiyan struck and is split between cells that were damaged and cells that did not experience winds strong enough to cause damage. Once more there is a fair bit of noise, though, with some spikes and troughs that are not easily explained based on conditions on the ground. The trough 10 months earlier could be due to typhoon Bopha, but that struck south of the Haiyan areas in early December (11 months prior to Haiyan) giving it a delayed reaction not seen for Haiyan and Rammasun. The peaks seen could be due to seasonality as they seem to appear in late spring.

As for Haiyan, the graph shows the devastating effect it had on nightlight values and similarly 8 months later, Rammasun would strike many of the same cells causing a second large drop in values. Interestingly there also seems to be a quick rebound in the months following the disasters, potentially due to the influx of aid and personnel. The impact seems to benefit the areas that were hit by the typhoon the most, implying that the aid mostly reached the people most in need. Overall, the results seem to indicate that the VIIRS data can be useful for a singular catastrophe such as Haiyan.

24

Figure 11: Philippines Typhoon coefficients with 95% confidence intervals


Moving on to a more national and the fixed regression analysis, different specifications of the regression were run, relating to the treatment of VIIRS values below 0, which were set to 0 if they were below a threshold of 0.3. Changing the threshold or running with original data did not affect the results.

Looking at Figure 11, the seemingly strong results from Figure 8 disappear when many events are included. Part of this could be due to many events happening within the 12‐month period leading to a confounding effect. The before seen measurement noise is also most likely a factor here.

25

Figure 12: Vietnam Typhoon coefficients with 95% confidence intervals


The results for Vietnam in Figure 12 are even more puzzling, as there is no significant effect until 7 months after the typhoon and the coefficients are very high compared to any other regression. This is most likely due to few events that would hit high light intensity cells.

V. Floods

The final natural hazard analyzed in this paper is floods. Floods are the most common type of disasters and it affects all the countries in this paper.

The modeling of floods can be done by remote sensing (Brakenridge and Anderson 2006) (Wu, Adler and Hong, et al. 2012) (Haq, et al. 2012) or through a combination of weather data and Geographic Information Systems (GIS) as for example in (Knebl, et al. 2005) (Asante, et al. 2007) (Dessu, et al. 2016). In this paper we will exclusively utilize the remote sensing method, due to the size of the areas that we model.

Even remote sensing methodology often uses an underlying hydrological model to model flow, see for example Wu et al. (2012) and Wu, Adler and Tian, et al. (2014), but in this paper only rainfall is utilized as a proxy for floods. The daily precipitation data used are from the 3‐hourly data set from the Tropical Rainfall Measurement Mission Project (TRMM), which is aggregated up to daily data. This has been used in several flood related publications before, such as Wu, Adler and Hong, et al. (2012), Wu, Adler and Tian, et al. (2014), Harris, et al. (2007), Rahman, et al. (2011) and Yuan, et al. (2019). Both Rahman, et al. (2011) and Yuan, et al. (2019) point to the TRMM data set being useful in flood modeling and in the case of Yuan, the TRMM

26

data set product 3B42RT 3 hour product gives the best results in a basket of 8 near real‐time rainfall products. The 3B42RT daily derived product is what is used in this paper.

To construct a flood proxy from the TRMM data, one has to define when a flood event is happening. In Wu et al. (2012) they propose four runoff based methods to define a flood threshold, and in addition Wu, Adler and Tian, et al. (2014) propose a slightly modified flood threshold definition with a point being flooded when:

𝑅 𝑃 σ 𝑎𝑛𝑑 𝑄 10𝑚 /𝑠

where 𝑅 is the routed runoff in millimeters, 𝑃 is the 95th percentile value and σ is the standard deviation of the routed runoff. 𝑄 is the discharge in cubic meters.

As the TRMM data do not provide runoff estimates, only rainfall amount, that will be used as a proxy for flooding in a cell. The modified equation is:

𝑅𝑎 𝑃 σ

where 𝑅𝑎 is the rainfall in millimeters, 𝑃 is the 95th percentile value and σ is the standard deviation of the rainfall.

Following the definition, one has a data set with daily flood data. To use this with the nightlight data it has to be aggregated up to monthly values. In this case, it was chosen to construct two variables as flood proxies, one with values for number of days and the other with values for total rainfall above the flood threshold for the month.

Once the flood data set is finished, a data set with the flood and nightlights data are made, ref Figure 4. With this combined data set, we do a case study of the 2017 floods in Southern Thailand and run the fixed effect regression:

𝐿 , 𝛽 Σ 𝛽 𝐹𝐷 , , 𝜃 𝑒 , ,

which is the same as in section II.c. As before, it gives the coefficients for the effect that excessive rainfall (floods) have on nightlights for the flood month and the 12 consecutive months.

Results.—The case study is the 2017 Southern Thailand Floods, which happened in the southern parts of Thailand during December 2016 and January 2017. These were the largest floods in over 30 years in these parts of Thailand. Malaysia was also affected, although to a lesser extent. There were almost 100 fatalities and the total costs was USD4 billion, making the floods fairly significant. The reason why these floods were chosen is threefold. Firstly, they lasted only 2 months, unlike the ones in 2013 that lasted roughly 4 months. Secondly, they

27

were isolated compared to other events. The 2014‐2015 floods would potentially show the 2013 floods. Finally, the floods caused significant economic damage.

One issue regarding the case study methodology for floods, was to define when a cell was affected or not, as almost all cells in an area would be affected if we chose the same flood estimator as above. To avoid this issue, a new “super flood” estimator was used. This estimator was set to 1 if the total rainfall in a month above our flood definition exceeded 500mm. Granted, this number was set rather arbitrarily, but it would capture a select set of especially impacted cells.

Figure 13: 2017 Southern Thailand Floods mean VIIRS nightlight values before and after for impacted and non‐impacted cells


Having imposed our super flood condition gave the graphs in Figure 13, which shows the means 12 months before and after January 2017. As can be seen, there seems to be a drop for the affected cells. However, this drop is well inside the regular noise level experienced during the period, throwing some doubts as to how strong the results are. Furthermore, given that some of the same cells also experienced floods in the month prior to January 2017, one might expect to see an effect in December 2016 as well, although we might capture an additive effect in January 2017.

Finally, for the fixed effect regression, different specifications of the regression were run. For floods, the key assumptions were related to:

1. Choosing a threshold for when a cell is flooded and not 2. Measuring flood in terms of days or rainfall sums 3. Treatment of VIIRS values below 0

28

Starting with the last point, once more it has been chosen to use a threshold of 0.3 absolute value as the base regression case. Choosing a different threshold did not seem to alter the results.

The first point is regarding how strict a measure one would like to have as a flood threshold. It was chosen to follow the assumption with 𝑃 σ, i.e. a standard deviation above the 95th percentile.

Finally, all regressions were run with a flood measurement both in terms of days of rainfall over the flood threshold and the sum of rainfall in millimeters over the flood threshold. The distinction between the two did not alter the results in a meaningful way and hence the total number of days is used as the coefficients are not tiny numbers.

Figure 14: Indonesia flood coefficients with 95% confidence intervals


Starting with Indonesia there is a 1% significant result in the month of the flood, which is strongly negative. This is potentially showing how most floods are small, local events where repairs are often carried out quickly after the event. Seeing the positive effect in month 8 after the flood is difficult to explain and is most likely due to a random effect in the data.

29

Figure 15: Myanmar flood coefficients with 95% confidence intervals


For Myanmar, the results are generally not significant as can be seen from Figure 15. Once more there is a significant effect a fairly long time after the event, which is probably again caused by a random effect in the data.

The Philippines (Figure 16) show a similar pattern to Indonesia, with negative significance in the month of the flood and then no further significance until much later. The overall effect is around ‐0.02 per day of flooding in the month, which is similar to Indonesia’s effect of ‐0.025.

Figure 16: Philippines flood coefficients with 95% confidence intervals


30

Figure 17 shows the results for Thailand, which show very little significance and only for 2 and 7 months after the floods. Once more it is hard to explain such effects. Likewise, Figure 18 depicts the same non‐significant results for Vietnam as for Thailand and Myanmar, and in Vietnam’s case there are no significant months.

Figure 17: Thailand flood coefficients with 95% confidence intervals


Figure 18: Vietnam flood coefficients with 95% confidence intervals


31

VI. Conclusion

Having used the VIIRS data as an economic proxy across modeled flood, earthquake and typhoon damage, the results were not convincing. The analysis in this paper shows a lack of significance across all hazard types and all countries. Furthermore, looking at specific cases for each hazard type only yielded strong results for typhoon Haiyan of the Philippines.

Overall, the cause for this is unknown, as multiple approaches were taken to try to correct the data and see if a connection between the nightlights and the disasters existed. As previously mentioned, the most likely culprit is the noise in the VIIRS data that is due to measurement issues and cloud cover. Another likely factor is that our sample of low‐income countries and the low correlation between monthly VIIRS and GDP is not a good fit for a broad analysis like this. These different factors might also explain why the authors found stronger results using the annual – and in most ways inferior ‐ DMSP data in Skoufias, Strobl and Tveit (2017).

That being said, the data and the methodology have shown earlier that there is merit in their use, with this paper simply demonstrating some of the limits of the data and that when they are used it is important to consider the temporal and spatial scale of analysis and carefully review both the data and the results to account for factors such as noise, cloud cover, light levels and urbanization.

As a follow‐up, it could be worthwhile looking at more specific cases to see if one could find an effect at a more local scale. One could also try different regression types, as this type of analysis could lend itself well to, for example, a difference‐in‐difference approach. Finally, with the addition of secondary data from other sources, such as for example daytime satellite images, vegetation images, national census data or similar, one could use machine learning methods to try to account for or correct some of the noise.

32

Bibliography

(GFDRR), Global Facility for Disaster Reduction and Recovery. 2019. 05. Accessed 05 2019. https://www.gfdrr.org/en/philippines.

(OCHA), UN Office for the Coordination of Humanitarian Affairs. 2016. Reliefweb. May 31. Accessed 05 2019. https://reliefweb.int/map/myanmar/myanmar‐natural‐disaster‐risks‐and‐past‐events‐31‐may‐2016.

Agency, Federal Emergency Management. 2006. "HAZUS‐MH MR2 Technical Manual." Washington, D.C.

Asante, Kwabena O., Rodrigues D. Macuacua, Guleid A. Artan, Ronald W. Lietzow, and James P. Verdin. 2007. "Developing a Flood Monitoring System From Remotely Sensed Data for the Limpopo Basin." IEEE Transactions on Geoscience and Remote Sensing (Institute of Electrical and Electronics Engineers (IEEE)) 45: 1709‐1714. doi:10.1109/tgrs.2006.883147.

Atkinson, Gail M., and SanLinn I. Kaka. 2007. "Relationships between Felt Intensity and Instrumental Ground Motion in the Central United States and California." Bulletin of the Seismological Society of America (Bulletin of the Seismological Society of America) 97: 497‐510. doi:10.1785/0120060154.

Atkinson, Gail Marie, and SanLinn I. Kaka. 2006. Relationships between felt intensity and instrumental ground motion for New Madrid ShakeMaps. Department of Earth Sciences, Carleton University.

Baragwanath, Kathryn, Ran Goldblatt, Gordon Hanson, and Amit K. Khandelwal. 2019. "Detecting urban markets with satellite imagery: An application to India." Journal of Urban Economics. doi:https://doi.org/10.1016/j.jue.2019.05.004.

Boose, Emery R., Kristen E. Chamberlin, and David R. Foster. 2001. "Landscape and Regional Impacts of Hurricanes in New England." Ecological Monographs (Wiley‐Blackwell) 71: 27‐48. doi:10.1890/0012‐9615(2001)071[0027:larioh]2.0.co;2.

Boose, Emery R., Mayra I. Serrano, and David R. Foster. 2004. "Landscape and regional impacts of hurricanes in Puerto Rico." Ecological Monographs 74:335–352. doi:http://dx.doi.org/10.1890/02‐4057.

Brakenridge, R., and E. Anderson. 2006. "MODIS‐based flood detection, mapping and measurement: The potential for operational hydrological applications." In Nato Science Series: IV: Earth and Environmental Sciences, edited by Jiri Marsalek, Gheorghe Stancalie and Gabor Balint, 1‐12. Springer Netherlands. doi:10.1007/1‐4020‐4902‐1_1.

Chen, Xi, and William Nordhaus. 2015. "A Test of the New VIIRS Lights Data Set: Population and Economic Output in Africa." Remote Sensing (MDPI AG) 7: 4937‐4947. doi:10.3390/rs70404937.

33

De Groeve, Tom, A. Annunziato, S. Gadenz, L. Vernaccini, A. Erberik, and T. Yilmaz. 2008. "Real‐time impact estimation of large earthquakes using USGS Shakemaps." Proceedings of IDRC2008, Davos, Switzerland.

Dessu, Shimelis Behailu, Abdulkarim Hussein Seid, Anteneh Z. Abiy, and Assefa M. Melesse. 2016. "Flood Forecasting and Stream Flow Simulation of the Upper Awash River Basin, Ethiopia Using Geospatial Stream Flow Model (GeoSFM)." In Landscape Dynamics, Soils and Hydrological Processes in Varied Climates, edited by Assefa M. Melesse and Wossenu Abtew, 367‐384. Cham: Springer International Publishing. doi:10.1007/978‐3‐319‐18787‐7_18.

Driscoll, John, and Aart Kraay. 1998. "Consistent Covariance Matrix Estimation With Spatially Dependent Panel Data." The Review of Economics and Statistics 80: 549‐560. http://EconPapers.repec.org/RePEc:tpr:restat:v:80:y:1998:i:4:p:549‐560.

Elvidge, Christopher D., Kimberly Baugh, Mikhail Zhizhin, Feng Chi Hsu, and Tilottama Ghosh. 2017. "VIIRS night‐time lights." International Journal of Remote Sensing (Taylor & Francis) 38: 5860‐5879. doi:10.1080/01431161.2017.1342050.

Emanuel, Kerry. 2011. "Global Warming Effects on U.S. Hurricane Damage." Wea. Climate Soc. (American Meteorological Society) 3: 261‐268. doi:10.1175/WCAS‐D‐11‐00007.1.

Facebook Connectivity Lab and Center for International Earth Science Information Network ‐ CIESIN ‐ Columbia University. n.d. High Resolution Settlement Layer (HRSL). Source imagery for HRSL © 2016 DigitalGlobe. Accessed 04 02, 2019. https://ciesin.columbia.edu/data/hrsl/.

Gillespie, Thomas W., Elizabeth Frankenberg, Kai Fung Chum, and Duncan Thomas. 2014. "Night‐time lights time series of tsunami damage, recovery, and economic metrics in Sumatra, Indonesia." Remote Sensing Letters (Taylor & Francis) 5: 286‐294. doi:10.1080/2150704X.2014.900205.

Haq, Mateeul, Memon Akhtar, Sher Muhammad, Siddiqi Paras, and Jillani Rahmatullah. 2012. "Techniques of Remote Sensing and GIS for flood monitoring and damage assessment: A case study of Sindh province, Pakistan." The Egyptian Journal of Remote Sensing and Space Science (Elsevier BV) 15: 135‐141. doi:10.1016/j.ejrs.2012.07.002.

Harris, Amanda, Sayma Rahman, Faisal Hossain, Lance Yarborough, Amvrossios Bagtzoglou, and Greg Easson. 2007. "Satellite‐based Flood Modeling Using TRMM‐based Rainfall Products." Sensors (MDPI AG) 7: 3416‐3427. doi:10.3390/s7123416.

Henderson, J. Vernon, Adam Storeygard, and David N. Weil. 2012. "Measuring Economic Growth from Outer Space." American Economic Review (American Economic Association) 102: 994‐1028. doi:10.1257/aer.102.2.994.

Hodler, Roland, and Paul A. Raschky. 2014. "Regional Favoritism." The Quarterly Journal of Economics 129: 995‐1033. doi:10.1093/qje/qju004.

34

Holland, Greg J. 1980. "An Analytic Model of the Wind and Pressure Profiles in Hurricanes." Mon. Wea. Rev., 108, 1212–1218. doi:http://dx.doi.org/10.1175/1520‐0493(1980)108<1212:AAMOTW>2.0.CO;2.

Hu, Yingyao, and Jiaxiong Yao. 2019. "Illuminating Economic Growth." unpublished.

INFORM. 2019. 05. Accessed 05 2019. http://www.inform‐index.org/Portals/0/Inform/2019/country_profiles/THA.pdf.

—. 2019. 05. Accessed 05 2019. http://www.inform‐index.org/Portals/0/Inform/2019/country_profiles/VNM.pdf.

International, GeoHazards, and United Nations Centre Regional Development. 2001. "Final report: Global Earthquake Safety Initiative (GESI) pilot project." Report, GHI, 86 p.

Jaiswal, K. S., and D. J. Wald. 2008. "Creating a global building inventory for earthquake loss assessment and risk management: U.S. Geological Survey Open‐File Report 2008‐1160." Tech. rep., USGS, 103 p.

Jean, N., M. Burke, M. Xie, W. M. Davis, D. B. Lobell, and S. Ermon. 2016. "Combining satellite imagery and machine learning to predict poverty." Science (American Association for the Advancement of Science (AAAS)) 353: 790‐794. doi:10.1126/science.aaf7894.

Knebl, M. R., Z.‐L. Yang, K. Hutchison, and D. R. Maidment. 2005. "Regional scale flood modeling using NEXRAD rainfall, GIS, and HEC‐HMS/RAS: a case study for the San Antonio River Basin Summer 2002 storm event." Journal of Environmental Management (Elsevier BV) 75: 325‐336. doi:10.1016/j.jenvman.2004.11.024.

Linkimer, Lepolt. 2007. "Relationship between peak ground acceleration and Modified Mercalli intensity in Costa Rica." Revista Geológica de América Central.

Michalopoulos, Stelios, and Elias Papaioannou. 2014. "National Institutions and Subnational Development in Africa." The Quarterly Journal of Economics 129: 151‐213. doi:10.1093/qje/qjt029.

Murphy, J. R., and L. J. OBrien. 1977. "The correlation of peak ground acceleration amplitude with seismic intensity and other physical parameters." Bulletin of the Seismological Society of America (Bulletin of the Seismological Society of America) 67: 877‐915. http://bssa.geoscienceworld.org/content/67/3/877.

National Disaster Management Agency, B. N. P. B. 2016. "DiBi database (Data and Information on Disaster in Indonesia)." http://dibi.bnpb.go.id/.

Rahman, Md. Mizanur, Dhyan Singh Arya, Narendra Kumar Goel, and Ashis Kumar Mitra. 2011. "Rainfall statistics evaluation of ECMWF model and TRMM data over Bangladesh for flood related studies." Meteorological Applications (Wiley) 19: 501‐512. doi:10.1002/met.293.

Skoufias, Emmanuel, Eric Strobl, and Thomas Tveit. 2017. "Natural disaster damage indices based on remotely sensed data: an application to Indonesia (English)." World Bank Policy Research Working Paper.

35

Strobl, Eric. 2012. "The economic growth impact of natural disasters in developing countries: Evidence from hurricane strikes in the Central American and Caribbean regions ." Journal of Development Economics 97: 130‐141. doi:http://dx.doi.org/10.1016/j.jdeveco.2010.12.002.

Uprety, Sirish, Xi Shao, Yalong Gu, and Changyong Cao. 2017. "Recent Improvements in S‐NPP VIIRS DNB Calibration."

Wald, David J., Vincent Quitoriano, Thomas H. Heaton, Hiroo Kanamori, Craig W. Scrivner, and C. Bruce Worden. 1999. "TriNet “ShakeMaps”: Rapid Generation of Peak Ground Motion and Intensity Maps for Earthquakes in Southern California." Earthquake Spectra 15: 537‐555. doi:10.1193/1.1586057.

Webster, P. J., V. E. Toma, and H.‐M. Kim. 2011. "Were the 2010 Pakistan floods predictable?" Geophysical Research Letters (American Geophysical Union (AGU)) 38: n/a‐‐n/a. doi:10.1029/2010gl046346.

World Bank. 2017. "South Asia Economic Focus, Fall 2017 : Growth Out of the Blue." Washington, DC. https://openknowledge.worldbank.org/handle/10986/28397 License: CC BY 3.0 IGO.

WorldPop . 2013. Viet Nam 100m Population. Alpha version 2010, 2015 and 2010 estimates of numbers of people per pixel (ppp) and people per hectare (pph), with national totals adjusted to match UN population division estimates (http://esa.un.org/wpp/) and remaining unadj. doi:10.5258/SOTON/WP00297.

Worldpop. 2016. Myanmar 100m Population. Alpha version 2010, 2015 and 2020 estimates of numbers of people per pixel (ppp) and people per hectare (pph), with national totals adjusted to match UN population division estimates (http://esa.un.org/wpp/) and remaining unadj. doi:10.5258/SOTON/WP00181.

Wu, Huan, Robert F. Adler, Yang Hong, Yudong Tian, and Fritz Policelli. 2012. "Evaluation of Global Flood Detection Using Satellite‐Based Rainfall and a Hydrologic Model." Journal of Hydrometeorology 13: 1268‐1284. doi:10.1175/JHM‐D‐11‐087.1.

Wu, Huan, Robert F. Adler, Yudong Tian, George J. Huffman, Hongyi Li, and JianJian Wang. 2014. "Real‐time global flood estimation using satellite‐based precipitation and a coupled land surface and routing model." Water Resources Research 50: 2693‐2717. doi:10.1002/2013WR014710.

Yuan, Fei, Limin Zhang, Khin Min Wun Soe, Liliang Ren, Chongxu Zhao, Yonghua Zhu, Shanhu Jiang, and Yi Liu. 2019. "Applications of TRMM‐ and GPM‐Era Multiple‐Satellite Precipitation Products for Flood Simulations at Sub‐Daily Scales in a Sparsely Gauged Watershed in Myanmar." Remote Sensing 11. doi:10.3390/rs11020140.

Zhao, Xizhi, Bailang Yu, Yan Liu, Shenjun Yao, Ting Lian, Liujia Chen, Chengshu Yang, Zuoqi Chen, and Jianping Wu. 2018. "NPP‐VIIRS DNB Daily Data in Natural Disaster Assessment: Evidence from Selected Case Studies." Remote Sensing (MDPI AG) 10: 1526. doi:10.3390/rs10101526.

Can We Rely on VIIRS Nightlights to Estimate the Short ... · Visible Infrared Imaging Radiometer...

Documents

Transcript of Can We Rely on VIIRS Nightlights to Estimate the Short ... · Visible Infrared Imaging Radiometer...