 Srečen sem, da je tudi praktykaličen, da poživamo praktykaliče. To je izgleda, da smo tudi praktykaliče poživamo praktykaliče. To je izgleda, da je izgleda, da poživamo praktykaliče. To je, da smo tudi praktykaliče, da smo tudi praktykaliče. Nr. inizaj, da smo tudi praktykaliče, da smo tudi praktykaliče. z vsej vsej situacije. Vsej vsej studi in v povali je, kaj sem vsej vsej, in sem vsej, da ne je nekaj vsej. Nekaj vsej vsej vsej vsej. Zato vsej polutanci nekaj vsej z vsej emisijanj. So we choose a radius of four kilometers in another situation. A major radius has to be chosen. This is an example of buffer, but another example is consider who is affected by this other environmental stress, very important, that is a motorway, ki je vsega vsega barfa. In in approach in approach in approach in approach could also be this one. To define a distance between between the population and the street and consider all people that are inside this buffer this linear buffer in this case as exposed to the street. So in this case the measurement is very different because we have a pollution that a pollutant that is emitted at the ground level so the dispersion is very little. And we often we usually talk about 100, 200, 300 meters, 500 meters in case of very heavy traffic in the streets. Of course, if we have information about residents and also workplace it's a good situation to consider this information together, we have an example of another incident in my region is in Parma city in which there is the the shape of the pollutant emitted by a dispersion model the population considering in our study on asthmatic adults so a very selected population and in this case we also consider by we also use a questionnaire to retrieve information about workplace and so you can see here this line is the distance between the residents and the workplace and to have the possibility to consider the two exposure is better that is obviously better. We have waited the time spent at home and the time spent in the workplace and we have measured the exposure to the incinerators in I far remember well 16 hours a day in the residents and the tower in the workplace and we found differences in exposure assessment respect to consider only address. I move quickly to the end of my presentation because this information are related also to the practicalities. An example of the importance of the introduction of GIS in exposure assessment studies in environmental epidemiology. We made a review of these methods related to studies of incinerators that is an example. We made this activity because we had an important project on incinerators and we want to understand how to calculate exposure. And in this review we consider three aspects of exposure intensity. So information about how much pollutant are emitted and how much pollutant go to the person. The spatial distribution, the scale and the spatial distribution of this pollutant and also the temporal variability that are the three aspects that together provide the best approach to exposure assessment. And we reviewed studies on incinerators starting from 1984 to 2013 and you can see here if considering the three aspects that I said before and going from the worst to the best you can see that with time there is significant improvement in the quality of the exposure assessment approach. And this improvement is mainly due to the use of GIS approach. So we can move to the practicalities. This last slide is about all GIS software that I know, commercial and free. Ok, thank you. So I will present now some practical example of how we have used GIS for exposure assessment in epidemiological studies and in anti-impact assessment. With the three practical examples one of these will be finished tomorrow with the calculation of the whole chain so the calculation of the health impact assessment due to this exposure. The first one is related to the study I said before. We made in our region a big study of all incinerators in the middle of the region. We have an incinerator for each province and also a total of nine incinerators and we made a study on also the environmental aspect but mainly the health effects of exposure to air pollution due to incinerators and health outcomes. And we published this work on epidemiology in 2013 with the main result of our studies that's related to reproductive outcomes and that we will show related to increasing exposure to incinerators. And in this slide I show the way we use to assess exposure. These are Emilia Romagna region that is in this part north of Italy and these are in red. The underlying shape file is related to the land user, the Korean Land Cover that is an European database that provides information about the use of land and the use of the territory and in red are reported the urban areas, this is Bologna, Modena, Reggio Media, Parma and so on. At the time of the study there was no incinerators in Parma but now we have a new incinerator and the example I showed you before was about Parma. We define a buffer of four kilometers around each incinerator and you can see that it's very close or overlapping the urban areas in different situations. So an overlapping also of different exposure because in urban areas you are exposed to traffic, to heating and very important environmental stress. So we have to consider the two aspects that are here, the dispersion model of incinerator and the dispersion model of all other sources in which you can see that the traffic lines are the main exposure factor. So we consider the buffer of four kilometers, we geocoded all addresses that fall into this buffer and we then construct a database of geocoded addresses with information of the name of the street and the X and Y coordinates and a database of population data for each address. We retrieve information about how many people and who live in each address and we reconstruct also retrospectively a database of a very desinential history. We consider the movement of population in time. On the other hand we have from a dispersion model we use the ADMS as I will show in the next slide and we construct a model for incinerators considering PM10 as the tracer of incinerator and all other sources that were traffic, heating, agricultural activities and industry. And we use Hanox as the main tracer of this kind of pollution. Inovoz, you cannot hear that for all other sources we are talking about microgram per cubic meter and here we are talking about nanograms per cubic meter. So a very high difference in three order of differences and more between the two kind of exposures. So this is the study design of the environmental study. We start from information about industry traffic, heating incinerators and also agriculture. And we made an emission map by means of ADMS. We have the dispersion model and the distribution of this dispersion model at the ground level. And to do this we use ADMS as I said before. ADMS is a very user-friendly window that allows a very easy use of the software. You can see that we can choose from between industrial sources, road sources and grid sources. The last one is related to information that is referred to a square. It is considered homogenous on a square of 500 meters or so on. We select industrial sources for incinerator. We provide information about the characteristics of incinerator. The emission factor that is provided by the owner and the control by our agency on emission at the stack level. And this is the output of ADMS. It is a text file with the concentration of different pollutants on a grid. This grid is defined by the first two columns with the coordinates. And so it is very easy to put this one in GIS file. This is the result. You can see that there is a regular grid, a regular base grid and the software had other points where the difference between values from a point to another is very high. It is called intelligent grid because there is more information where there is more heterogeneity in the concentration. So here you can see the roads. This is the motorway that I showed you before in the example. And regarding the exposure of population, what can we do with this information? We put in these triangles the address of each person and then we have to move to exposure assessment to this, related to this map that is in this case the map of all other sources. How can we do that? If we see to this address, for example, we have to choose from different methods. One of these is the nearest monitor. Otherwise we can choose to construct a map by interpolation. And this was our choice to have a more homogeneous distribution of information and arc map in this case, but every GIS software can provide a tool to do this. In arc map is geostatistical analysis that allows us to make calculations about inverse distance weight, creaking and so on. We choose creaking in our exercise, but with this very rich information, very rich variability of data, the result is very similar between creaking and inverse distance weight. And this is the map that is provided by the creaking application and in this way is very easy now to assess exposure to each address simply making the point in polygon analysis that I showed you before. And to consider what is the number that is under each triangle and assess exposure in this way. We can also add other information. In this case, this is the information about census block. Census block provides us information about socioeconomic status and so with a point of polygon analysis, again, we can provide information about socioeconomic status to all our court with the assumption that all people living in the same census block has the same socioeconomic status. That is an approximation, of course, but it is the most defined information about the economic status that we have. So this is an example of what I showed you before in the theory part about the overview analysis. Another example that is again on waste about landfills in Europe. This is an example that I put in the presentation for two reasons. One is an example of how to use very large information, already available over all Europe to calculate an impact due to environmental stress that can be compared between member states, between countries. And the other reason is that this example is complete for our purpose because there is also the part of calculation of attributable cases that we will make tomorrow after the lesson about the concentration response function and the integrated approach to adding impact assessment. So we will be able tomorrow to complete the exercise hopefully with an XR5 together. So for now we concentrate to the first part that the GIS approach to exposure assessment. What are the information we use? We want to calculate how many people were exposed to landfills in Europe and considering the exposure with the simpler approach we can use at this distance because we consider a lot of landfills, 1000 or perhaps 100 or perhaps 1000 of landfills. And so we don't have the capability information about the dispersion model of all these plants. And so we consider distance as a property of exposure. We need the location of plants. The location of plants was provided by the European Pollutant Release and Transfer Registry, that is an initiative of European Environmental Agency, a very good database that allows us to geocode, geolocalize all landfills in Europe. And the other aspect was the population. And the population were provided, again by the European Environmental Agency, by a population database that is a very big raster file with information on density population across Europe, across a large part of Europe. You can see that not all states are reported here, but it's a good number of states. And I mentioned that we will return tomorrow also the head for all database that Massimo introduced in his lecture this morning. We used this database to consider the background incidents of mortality and morbidity of population divided by country. So EPRTR register is a very, very powerful register that is regulated by European law. And he reported from 2007 to 2014 all the information by EU member states and then Iceland, Litterstein, Norway, Serbia and Switzerland. And as you can see in the slide, there are a huge number of information and polluted registry from industry reported, more than 30,000 facilities in 32 countries. This is the representation in Google Earth, the format of the file provided by the European Agency is the CAP KMS file that can be visualized in Google Earth. And this is a representation of the sector 5, 5.D. The sector 5 is a waste management premises, the subsection and the sector 5.D is landfills. And in this slide you see all landfills that are geocoded in Europe. This is the table of the landfills we have considered. There is this part that don't allow us to have the total, but I think it's more than 1,500, I think. And on the other hand, we have information of population by this, other, very useful, I don't know if it's a registry, it's information, all again provided by European Environmental Agency that is a big raster file with a resolution of 100 meters in its square there is information about density population, population density. And this is the graphical representation of this raster file. And in this way we can put the population, overlap the population to the location of the landfills. This is a KMS zeta file. I show you some example in ArcMap, but the name of the function is the same in almost all software. So in ArcMap there is a tool to convert KMS zeta file to layer, that is the shape file used in ArcMap, but also in KUGIS and other software. So we import this information in ArcMap and this is the result with the mapping of all these landfills. For each landfills there is a zoom on the north Italy. You can see here the Garda Lake and here the boundary of Italy. You can see here we construct for each point with the instruments of buffer, different buffers and this is another zoom. In this slide you see the buffer and the information about the population. These are the single 100 meters square with information on population density. So starting from these two information, the location, the buffer and this raster, we were able to calculate the population exposed to each of the considered landfills. Another example that I show you is a free package that is GMA that stands for Geospatial Modeling Environment. This is a free package that you can download from the site. It's the only problem of this free software that is that work only if ArcMap is installed on your computer. So it's a very powerful tool because simplify a lot of functions in ArcMap that are very difficult to calculate but require the engine of this software. In the website there is written that they are developing this software for other but it's a year that I don't check but a year ago there was no improvement in the sense. But it's a very good tool to make an intersection and overlapping in a very easy way. This is a window that we use to calculate this population. All this calculation could be made only with ArcMap or QG. But it's more difficult, this is a very easy way. And I show also some other suggestions about the calculation of population in industry, in insulator and so on. And indicating the name of the tool you have to use in the different steps. QG is ArcMap and I don't know also if grass and other. Name proximity analysis tool, the way to use to calculate buffers. And this is the window in ArcMap in which you can easily import the point sources from which calculate buffers. The unit 100 meters or so on to calculate to define buffer. In this windows I don't remember if no, there is another window. In multiple buffers you can calculate simultaneously different buffers. And in this case, this example is a three kilometer buffer. And to calculate exposit population, we can have information about the square like in the environmental agency database, raster file, or if you are in your country or in a single situation, you can also have information about sensors block data. In Italy when we made this calculation only for Italy, we often usually use this information about sensors block. In each sensors block we have the total number of population that live inside. And with the approximation that this population is uniformly distributed into the buffer, we can calculate how many people live in the green part around this plant. To be more precise, we have to choose the criteria to consider the sensors blocks that overlap the boundary of the buffer. And how to make the estimation of people living in this type of sensors block, a way is to calculate exactly the number of population in the assumption of the uniform distribution into the sensors block. And this can be made by two steps, the intersect function of our map or could just intersect between buffer and sensors block. The calculation of the area that is a function that is built in the GIS file to calculate, for example, this area starting from information of the whole area and to apply a simple formula to calculate which population, how many people live in this single part. This is the function intersect, OK. And this exercise tomorrow we will complete tomorrow with the 18-pack assessment calculation, considering also the outcomes, at least a couple of outcomes that literature suggests in this situation. The last slide, the last two slides are related to the... I think, no, there is more than one, OK. The situation about the construction of land use relation modelling in case extensively talk this morning about this powerful tool to assess exposure to environmental stressor, in particular in urban areas. And I took this part of the slide from case that is the list of information of predictor variables that has to be built for the implement of the relation model for the land use relation model. All these information are built by GIS approaches. And in this graph you can see the way in which are considered the instrument, the four or five commands that I show you during this lecture are sufficient to calculate the number of roads within a buffer, the number of roads or the number of major roads, the area of the use of the territory in a buffer. So information about, I can see here, land use, the first one, you can calculate this area in the same way I showed you before, and decide into this buffer what is the main use of the territory in the buffer. Or you can consider the number of industry that falls into a buffer. You can multiply to a value of emission, if you have. All these information are calculated with the simple instrument that I show you in this presentation. And all these variables are calculated in this way. This approach is used also in other situations. This is an example we made to compare GIS approach in recidential exposure compared to self-reported exposure by questioner. And there were interesting situations about the recall bias, the demonstration of the existing recall bias or differences between people living very close and people living far from industries and so on. And to close this presentation, when we have all these aspects, we can apply all these approaches also when we have uniformly distributed pollution, such as when we have to calculate the environmental burden of disease or rating of assessment of PM 2.5 overall state or overall the world as some example in the lecture before as shown. Also in this case, we have a map in a raster file on other sources and the underlying population that allow us to make all these calculations in the same way I showed you before. This is an example from the VAS project that is a rating of assessment calculation of pollution in Italy that Carla Francesco and colleagues have made some years ago, in which there is an example of a grid in a 4 km grid of pollution overlapping information about population by census block. And this is perhaps the worst situation in which a census block overlapped four different squares, but it is possible with the approaches that we said before. To calculate how many people are exposed to these values, how many to this, how many to this, how many to this. So in the iterative process, all these calculations are possible to define the more suitable way to assess exposure also to pollution overall. Thank you. Thank you very much. And thank you for updating the presentation so fast for the tumorization. And is there any question? OK. Oh, OK. Thank you so much for your nice presentation. So what I'm thinking about that dispersion of pollutants is always affected by land use land cover as well as topography. And especially in urban areas we have high rise buildings. So how this model is taking care of the morphology of the settlement as well as land use land cover present in the area? How this model is taking into account the topography and... Yeah, land use land cover, topography, as well as morphology of the urban setup. How kind of model? Which kind of model? This dispersion, pollution and dispersion. OK. There is no more. There is... You can have information on... No, no, it's OK. A layer in a GIS file about topography that is very important in some situation. In the... I don't know if ESCAPE project is mentioned and is being mentioned during these days. OK. In ESCAPE project land use relation model was made for different cities in Europe and in situations like Rome, for example, in which the topography is relevant. This was a variable that was determined in the characterization of pollution. Not in my city, for sure, because I live in Povil in this situation, it's not this one. So, both dispersion model and land use relation model can take into account properly this information. I don't know if this was your question. No, about that use of... The importance of those variables in dispersion model, in your case. Yeah, yeah, OK. Thank you so much. Yeah. OK. Thank you very much. Sorry for the presentation. If the local agency does not have, maybe for example, the vector or the raster data format of the area you are interested in studying, what is the source of data for the games? Without information about local environmental agency? Yes. Yeah. Perhaps... I have shown a slide in which there is information available. For dispersion model, it is crucial to have good information about emissions. And emissions can be provided by local agency, but also by the owner. There are those that... Emission data, yes. The local refinery can produce. But for example, the land use, land cover or the terrain. Those, yes. If I put the exact name, but... Yeah. The existing maps are related to... When we talk about holography, topography, there are European data set that can provide this information. And I know information, particularly at the European level, but I think that it's extensible to other situation. In Europe, there is a project that is covering and covering, that every ten years or perhaps also every five years with some improvement, provide information about land use in a square, but a very fine resolution. And also there are information about a main criticality about the availability information related to traffic and to roads, but in particular to traffic, because the way to calculate traffic difference from city to city, from region to region, from state to state. But if we refer only to the network of roads, we have the Aero Street Fire, for example, that is the map of the old roads that are in Europe with the classification of roads, that classification of roads is roughly an indicator of traffic. And so there is a very high availability of information of this in this way. But only Europe also elsewhere. This is a good question. Yes, I don't know in the other situation. I don't know if, I'm having a look now, but I don't know if there is the specific information you might want, but Google Earth Engine is available for, it has global data and it's free to use. And it might be useful when we're missing data, there's no data provided by the government or by other studies. Yes, and also other initiatives that free open street map, but I don't use open street map because I have information about in Italy and Europe available information, but it's a free website with information all over the world about street, I don't know the availability and the level of uncertainty related to this information, but it's a website that is called Open Street Map. And I think that there are also other... There is a website called geeta.org. They give a lot of data, raster as well as vector globally, so they can download for their country, about the topography, about the land use land cover, about population, a lot of data is available at geeta.org. Or diva.org is another website, where you can download a lot of GIS software as well as data. And government of India is also providing data through Bhuvan portal, so you can have a lot of data from Bhuvan portal. Excuse me, just I have a comment for Mr. Biblap, when you're working on air dispersion model, you need two kind of file, terrain elevation data and land use data. And you can go to the website of USGS, United States of Geography, and enter the longitude and latitude of your area and they will give you this kind of file for all the world. USGS? Yes, yes, yes, yes. Thank you. I also wanted to add a comment to what friends told. I think that OpenStreetMap is very good in such cases for our countries, because people, volunteer people produce data, shape files, and they upload it. And also, and they are mostly very useful, you can have many versions of the data and you can select which one is more exact and use it. And also USGS also have mostly in images, raster image, you can use USGS. And also if you search in the internet, some people who provided the data mostly upload it for free or for some small yes, and you can use it. It is a very good point in these days. So I think we do not have any shortage of data. Maybe for your specific, what is specific for your field of study, you have to digit some data, but for other popular data of land is land cover or a street or river, it's free and you can use it. Excuse me. For health impact assessment, we need emission data, yes. But if we don't have these emission data, we have only concentration. In our country, the emission data is not available for everyone, but the concentration, some places it's available. So how I can convert from concentration to emission to make these health impact assessment. Another question, can I use the data of climate change scenario to make health impact assessment for the future? Thank you. Regarding the emission concentration, the health impact assessment requires information about concentrations. So if you have concentrations, you can make an exercise of health impact assessment, because evidence from literature are related to exposure to concentration of pollutants. I have talked about emissions data in the case of application of a dispersion model of a point source. In this case, it's crucial to have emission data, but if you have a good network or good information about concentration, it's okay. And regarding, and I don't remember the last... It's a separated station. Maybe, for example, in my authority, we have available five stations, but it's one in the desert outside the boundary of Egypt, one on Cairo, one in different places. Not many stations. Yes, it's the same argument of tomorrow morning about the number of the location with the information. It depends on which model you have to apply in case of end-use creation model. As case said, less than 20 points is not useful. And also you have to consider the distribution in the area. And if you are interested in health impact assessment of short-term aspects, so the variation in time, you can use information from a sparse monitoring data, but you consider only the variation in time in days, so you have the capability of predict health outcomes related to short-term effects. I just want to comment what she was saying, what if we don't have data information about the emissions, but just the concentration, because now we know that one of the main points in health impact assessment is to provide scenarios and so evaluate the health effects when something is changing. What if we cannot say anything about the emission because we don't have the data? You can work on the concentrations. For example, in the example in the Andrea show, the Italian air pollution health impact assessment study, we sometimes the scenarios are asked from the politicians, the decision makers, sometimes they don't know, and so we decided to create our own scenarios, working on the concentration, I mean, just saying, what if we can reduce of 20% the concentration, what will be the effect on the health? What if we cut the concentration and keep just the limit by law? Is something that you can do cell by cell, simply the calculation, you can just reduce the concentration and evaluate the relative impact, so you can still do some things and assumptions, it's a scenario, theoretical scenarios, but still you can do some, so even if you don't know the exact amount of emissions, which is usually the case also in Italy, I mean, there is no common to have such information. And this is very useful policy makers, very often you say we go to the WHO's level, we can earn, save, et cetera. I would like to add that since this morning we have heard a lot, it probably is important to underline that the lure model can be used, even if you don't have any monitoring station, or you have some campaign, you assess, tomorrow I think you will go back to that, but you can assess air pollution, and then from that you can extrapolate and have some basic data. Obviously there are advantages and disadvantages, but in a situation where really you don't have anything, this is a good start, this was adopted in escape, where in the towns where they didn't have anything and as a given a nice result you publish a lot on that. For example, now in Rome we, I mean, not as citizens groups, organized groups, they decided to monitor the air pollution because of traffic, so they buy this monitor, it's quite cheap for, I don't know how many euros, but quite cheap, so they put the monitors for one month, and so now we have 250 points in Rome, just with one campaign, of course it would be better to have one in one season, and one in the opposite season, to take into account the variability, the seasonality, the variation in seasons, but still you can do something, and we are now conducting, build this lure to estimate concentration in other points of the city, taking into account all the other variables that we have, traffic flows, population density, distance from main road, the more the better, I think. That's really useful, in France we have cars, and now doing that, and the individuals, so cars go everywhere, you have a border, that's in Paris the mayor, I need to go that supported this action to better monitor pollution, because we have only 22 monitor stations, only three assess PM 2.5, zero, the ultra-five, so boo, yes, and where it's really very much polluted, they have taken out the monitor station to reduce the cost, because, you know, this is another criteria. Also to estimate the production or the emission, you can use an inverse model, if you have a dispersion model, you can write a procedure, in the direct model you specify the emission, and then you calculate the concentration, and the inverse model is doing, sorry, you take the concentration and you derive the emissions, but it's the arms work to do, it's not trivial to do that. OK, any other question? So may I suggest that for the person who shared some information about ways to make a low-cost GIS, I don't know if I got it right, but maybe you could share this information on the questioner, so that everybody knows where to find that, and for tomorrow we've decided that we should move, start a little later, 8.45, instead of 8.30, so that we are able to have breakfast and to take the shuttle. 8.45. 8.45, yes, and tomorrow is holiday day, it's a national day of Italy, so the shuttle bus, if you have to take the bus, it's like a Sunday for how you call it, the time schedule. OK, are we done? Yes, I guess. So now there is the follow-up of the exploration of the posters. OK, there are the two operators that should start. Kala, are you leading a two operator for the poster? OK, so we meet at the poster, the one, I think we started some of it. I think yesterday works in a nice way, what do you think? I mean, I would like to listen from other students, so we can continue our tour or exchange, I don't know. OK, so it will be me, Francesco, Andrea, Isabella, Jean-François. OK, Jean-François. Yes, bye.