These are a single variable or combination of variables that provide information on an underlying concept. We are particularly interested in indicators that can be created from a census and refer to relevant underlying concepts like exposure, adaptive capacity, human capital, social capital, as examples.
by Amy Kracker Selzer, College of William and Mary
On the “methods” page, I discussed the creation of a village-level risk hazard index. This index aggregates information about the geographic distribution of risk for flooding, landslide, and drought to create a mean aggregate risk value for each village in the Semarang study area. This post further discusses the creation of this index as well as its potential uses and limitations.
As described by Sainan in her post “Digitize a JPG picture,” the hazard data that we were able to obtain for our study area came in the form of a jpeg that visualized variation in risk levels for specific environmental hazards but was not in a format that allowed for calculations within GIS software. In order to have this functionality, each jpeg needed to be converted into a shapefile (see Sainan’s post for details on how this was done). This conversion outputted one polygon shapefile for each hazard risk (flooding, landslide, and drought) that allowed for the visualization of levels of risk across the study area and calculations between layers. Each of these layers is individually useful for examining the geographic distribution of a specific environmental risk within Semarang. However, in addition to understanding the geography of each individual risk, we were interested in examining the spatial distribution of aggregate risk that encompassed all three hazards. Therefore, we needed to combine these three risk layers into a single indicator. Each hazard risk shapefile was standardized to represent levels of risk on a scale of 1 to 7, with 1 being the lowest level and 7 being the highest (see my post “creating a village risk hazard index” in the methods section for more details). These risk values then needed to be summed for each location to create a geographically specific aggregate measure of risk. This first required converting the polygon hazard risk files to raster files. Next, the raster files were overlaid on top of one another in ArcGIS and specific pixel based summations of risk levels for each hazard type were made, allowing for the creation of a continuous surface representing the geographic distribution of aggregate risk. In addition to examining this distribution, we wanted to compare this aggregate risk measure with village-level census data. Therefore, each village needed to be assigned an average risk value. This was done by overlaying the village boundaries on the aggregate risk raster file and calculating a mean risk level within each village using zonal statistics.
The use of GIS software to calculate risk across various hazards provides an opportunity to examine the geographic distribution of aggregate risk and comparing this information with data from other geographic scales. When using the resulting measure to make assessments about hazard vulnerability, however, it does assume that the consequences of each hazard type are likely to be similar within that space. If however, the consequences of flooding are likely to be greater than the consequences of landslides, flooding risk may have to be differentially weighted prior to summing the risk values for each location. It is difficult to suggest specific weights or weighting methodologies as these will need to be specific to the research questions being posed and the data being used. For our purposes, we determined that the impacts of each hazard type are likely to be relatively similar and no weights were used.
Another important issue to consider when developing this kind of indicator is the way in which the level of analysis impacts upon the data. The aggregated hazard risk raster file described above is able to display a continuous surface of risk at the level of small pixels. Therefore, it provides the most accurate representation of information about risk at the smallest scale. However, we were particularly interested in comparing the hazard risk information with village level demographic data. In order to do this we needed a village level indicator of risk. As described above, this was done by using zonal statistics to calculate a mean village risk level. When aggregating this risk information to relatively small polygons such as villages, some geographic variation is lost but the information remains relatively accurate. However, if a mean risk level were to be calculated for large polygons based on this aggregate raster file, much of the variation within the geography would have been lost and the polygon-level value might not accurately reflect the variation in risk within that space.
/indicators/upload_files?prefix=20170627213728/components/com_communityanswers/assets/uploader/plupload.flash.swf/components/com_communityanswers/assets/uploader/plupload.silverlight.xappdf,doc,docx,xls,xlsx,odt,rtf,txt,odf,csv,dbf,odp,uop,pptx,ppt,pps,ppsx,eps,gif,jpg,jpeg,jpe,png,ps,svg,tif,tiff,bmp,cdfClick here/indicators/verify_captchaContribute Your ResponseCancelYesNoError!ConfirmSuccess!Click here/indicators/subscribe_qn/38-developing-and-using-a-climate-risk-hazard-indicator/indicators/unsubscribe_qn/38-developing-and-using-a-climate-risk-hazard-indicatorYou have successfully subscribed to this topic. You should now receive emails when new respones are posted.You have successfully unsubscribed to this topic.Subscribe To TopicUnsubscribe From Topic/indicators/get_answerReport ThisSubmitInformation!