 Welcome to the NPTEL course on remote sensing and GIS for rural development. This is week 10 lecture 2. In weeks 9 and 10, we are looking at specific remote sensing data that can be used for understanding land available for rural development in priority agriculture and how expansions can happen without compromising the natural resources. On that note, we were looking at LULC change and one of the key changes for LULC is multi-cropping which includes rubby and zide cropping. If it is just monsoon crops then or rain-fed crops then we won't have much issues. However, if the cropping is increased, the cropping cycles are increased in the monsoon and zide seasons, then there is tremendous pressure on the land resources and water resources. This directly impacts the further development in rural regions. So therefore, there is a need for evaluating the suitable land for agriculture and development. The development could be industrial development, housing development, etc. So, there is a need for using data to understand potential areas for development and there is also a need to use data for monitoring. Suppose a development happens without planning, we need to understand if the development has an impact both positive and negative on the rural regions. I am talking about the natural resources and potentially the population. While these are not straightforward answers, many a time there is less data collected for these exercises. Therefore, there is a need to collect other data and we were looking at data mining proxy data approaches of which remote sensing and GIS platforms help a lot. So, we will continue our discussion using remote sensing data for such rural development scenarios in this lecture. At the end of 10 week lecture one, we notice that crop statistics is an important information that is needed for rural development for reducing the impact of climate change on farmers and rural communities and also for planning for future scenarios. In that case, there should be data for statistics and we understood that there are a lot of times there is data lag or insufficient data to make a conclusive evidence. So, therefore, we will be looking at using secondary data remote sensing data for this purpose. So, let us go in as we discussed to explain what is synergized mapping. So, synergized mapping as it is being trademarked is bringing multi-source data and multi-disciplinary data together so that scientifically validated management plans can be developed or frameworks can be developed that can be readily applied to rural development scenarios. Let us just take the case of the Rabhi and Non-Monsoon crop increase in rural regions and how it impacts or how crop statistics is needed. So, as we mentioned, if you just have government agencies and if they are just collecting data by surveys and some ground proofing exercises, there is a big delay in the data that comes into the network and therefore, there is unsustainable agricultural practices, the benefits are not shared equally and the economic growth is reduced. Moreover, the risk management is breached because there is no early warning system or reflection of what is happening in the ground. Let us say for example, there is a cyclone, maybe there is an early warning system to predict the cyclone but there is no system that looks at the post cyclone analysis on crops. There is no mandatory and there is data expenses. On this note, we are looking at using remote sensing data for filling that gap. Let us see how that goes. So, in synthesized mapping as you could see, there is multiple institutions, agencies that take part in data collection using multiple sources and multiple methods. So, you could see governments, institutions, farmers who are the key stakeholders or rural communities and NGOs who handhold the farmers and rural communities, all of them collect data and they may be trained in a particular data set and they may not be talking to each other in terms of let us say the health institutions are taking health variables, water quality variables, whereas farmers are only taking water level variables. So, this is how we can have different players who are experts in certain disciplines but overall theme is rural development. Under that, there could be multiple teams such as rural education, rural healthcare, cropping, farm allied services. All of these can be mentioned and monitored and managed by different agencies. So, government always collects data whether it is enough or not is a second question but it always collects data through surveys mostly surveys and some measurements. So, let us say government is taking that data and then we have institutions and NGOs who are slightly advanced in technologies. Institutions include academic institutions like IIT Bombay where I work and that could include data capturing and data analysis using remote sensing products and GIS platforms. So, we have institutions who can do remote sensing and smart tech. We have governments who are doing surveys and ground proofing. The farmers themselves can provide data back to the system using smart tech like mobile phones or just an SMS or a WhatsApp message saying that if the water level is there, what is the water level or what crops they are growing. These are all important. See the crops, they could change within a week. If it is not growing properly, if some calamity happens or animal grazing happens, they just rip it off and put a new vegetation if needed and that is where the ground up crowd sourcing data plays a vital role. So, we have the farmer's data that can come in as not surveys because that government actually asks at maybe once every three months. Farmers can give every week or even daily a WhatsApp message or a WhatsApp image of what is the water level. And then we have smart tech and surveys combined as farmers. NGOs and institutions can be a little bit higher in technology and so for example, I was working with three NGOs before I joined IIT Bombay and the technologies they have is really novel and cutting edge. So, this helps in bridging the gap of issues if any like in terms of new technologies for bridging higher special resolution and temporal resolution data. So, NGOs and institutions can help. Farmers can help provide ground up data and government provides whatever data can they collect. All of them can come together. All of them are different disciplines, can be different disciplines. For example, the government can have a policy angle whereas the farmers can have the actual angle of the ground and they may be using multiple different tools, remote sensing smart tech surveys and at multiple spatial and temporal frequencies. At the end of the day, all of them come to a one database and is being analyzed for rural development that is called synergized mapping. So, we do have satellites and drones in the remote sensing very, very I will pick up what are some examples of these tools. So, in the remote sensing we do use open source satellites like NASA's MODIS, Landsat, our ISROS resource set and drones are there from Agricultural University mapping and exercises. Then we have crowdsourcing data from farmers. The farmers can give individual data or also NGO trained data through farmer networks and communities and also there are multiple mapping communities that provide data. For example, OASM is a very good community that provides a lot of data for updating the attributes online. So, in that network, let's have some examples of this synergized mapping. What we will be showing now is some data sets that have been created using the synergized mapping framework, especially the OASM data in the weeks to come. So, we have understood that NDVI can be an indicator for crop health monitoring and crop growth wherein water application fertilizer application can be given. So, let us now look at the different NDVI products that are available and different platforms. So, ISROS has its own ISROS-driven data and also NASA data-driven products in ISROS website. And then we have the Google Earth Engine data catalog. The NASA's, USGS portals also have different images and accessibility is pretty straightforward using the Earth Explorer. And then we have the European Agency Sentinel Hub. So, I am giving the key ones which are highly used by scientific communities and researchers in the world. So, let's look at the first link which is the ISROB1. I will be sharing my screen now. We look at all these examples. I will show you the methods to go and look at the data and download the data and use it for your exercises. We have already showed how to download data from Earth Explorer, data engine, etc. It's the same format and there are multiple tutorials just to download data but less on applications of this data for a particular cause where we have picked rural development as the cause and we will be looking at the different products. If you look here, the products are different not because of the algorithm. The algorithm is the same NIR minus red by NIR plus red. We are not changing that formula but the resolution, the spatial temporal resolution of the data could be different depending on the instrument that is used and the data availability, cloud source coverage, etc. is different. So, we will be taking those data that are very helpful for the given location and we will see how that can be used widely across the rural regions in India. So, let us go ahead with the first platform. So, you could see that the first link, let me copy and paste. So, it is going out of the slides. So, I am going to open the first one where we will be discussing about the booban data set. So, I will be opening the booban link because what is happening sometimes is the link to the data set might change and get updated. So, let me share the boobans web page. So, here you have the boobans web page and what you have here is the third, sometimes this open data archive or thematic maps as shown below, you could use thematic services. I will click it is the same or you can open the open data archive. So, thematic opens two different portals. So, as I said the first one from here just visualize and download, we will be looking at the fourth tab which is the open data archive for accessing the booban and dvi data. So, I am just going to click it. There you have the booban data set and also if you come down, you will see thematic services. I will be using my pointer now. So, we will be having the thematic services. I am going to open that too. So, first let us go to the NRAC open EO data which we opened first and in that there is theme and products. So, first is satellite sensor. You by reading you know that OCM2SAT has been used widely for NDVI calculations. However, when you look at it, there has been some upgradation doing better for better facilities. So, as I said, even though I gave you the direct link to go and look at the thematic layers, suppose in this NPTEL course is run in a later time when this website does not work, it is always better to go to the parent website which is booban and from booban you can easily identify where it is. You can also type booban NDVI as I have searched before this class and you could see that there are multiple links to open booban data. So, for example, if you open this, it again reroutes it back to the same products as I mentioned. Okay. So, let us go to the first list and by satellite if you access, it gets difficult. So, it is better because sometimes as it is said, OCM2SAT is not giving. So, then you can go to resource set and then see what resources are being met. But we are going to go at theme and products because we want a product out. The raw data will have NIR and red, but we do not want to calculate it because it is already there. Why do you have to calculate it when it is already there? So, we are going to use the boobans data products and you can see land and terrain is there, land vegetation is there. So, of the indicators for agricultural and rural development, we have mentioned that the NDVI is very key and then we have the vegetation fraction cover is also there. But let us go to NDVI if available in the land vegetation. So, I have searched the select theme is land vegetation. So, let us go by terrain. Terrain will just give you what are the versions of snow cover, albedo, DEM, the digital elevation model, etc. We are going to do that. We are going to go to land vegetation and in the land vegetation, we can see that there are the four different parameters of which two data products are there. As I mentioned, the first one is NDVI. So, you have the filter normalize difference which is index. So, it is a filtered NDVI and then we have the NDVI global coverage. I am just going to click to see that, okay, it goes global. And then we have the local coverage where we are just going to look at India. OCM 2, again OCM 2 is the satellite that has been used. But if you go to satellite sensor and do OCM, it would not show because it is saying it is updated. Here it is coming up. And then the last one is the vegetation fraction which we can take all OCM. So, let us go to the first NDVI local coverage. And then we can, as I said, look at the brochure of what this has been done. These are the satellites of OceanSat to the satellite, the range, et cetera, et cetera. We will anyway look at the metadata when we download the data. So, all this can be read for interest. And then we have the technical documentation add-and-down, which is an addition to the technical data site. So, it says NDVI and vegetation fraction, how they did it, what is the methodologies, et cetera. It is a 15-day NDVI composites. So, every 15 days, the data was collected. And then the cloud cover was removed, but analyzed, generated using earlier methods and MVC method where the cloud cover has been removed and or negotiated. So, we have the range, as I said, in the previous example, 1 to minus 1. 1 is vegetation, whereas minus 1 is water bodies. And we can have them. So, the last is the technical document. It is always important to go through these documents. And it says, almost sometimes it is duplicated, but we will just see if something we need to be careful about. For example, the spatial resolution is around 1 km, right, so 1000 meters. And the georeference coordinates are given, data processing, how they did it, and the NDVI function, what they used. So, as I said, NIR minus red, NIR plus red, and that is what we also used for our index. And then the vegetation fraction is NDVI minus NDVI naught, and NDVI infinity minus NDVI naught. So, if you look into what these NDVI naught and just NDVI is, you will get more information about, is it reflecting the curry-frabi or zine or double cropping. And then, so how do you determine the Instagram of vSpeed fixers was then used to determine NDVI naught and NDVI infinity. So, again, these products are kind of less used compared to NDVI. So, I am going to first show the NDVI product, and then we can have a call. So, look at this NDVI product, and then we can go to individualized products or year-wise. If you click year-wise, you'll first know how many years are available. Okay, so from 2011 to 2021, so approximately 10 years, 11 years of data is available on Bhuvan. But you can go to 2021, and then you can just quickly see which are the maps. If even if you do individual products, the same thing, click the calendar, click the calendar, then the year comes and the individual product comes until December. You want to go to the previous year, you can click the previous year, and there we go. So, let's go to the recent, most recent one, which is December. And then you can click metadata for December. The document has the full forms, but the most important ones, we will quickly look at what is the addition? What is the data? It is for NDVI and VF calculation. The coordinate reference system is DCS, WGS 1984, which is good, which is what we're using throughout this exercise and lectures. Name of the satellite is important, so ocean sat. And then in what format it comes, it comes in geotiff, the spatial resolution degrees, and it has been given approximately as one kilometer. Okay, number of bands is just one because this is a product. It's not like a satellite raw image. If you go to a satellite raw image, you'll see like number of bands, and that has been calculated to get this data product. Okay, it has been rectified. Some, some of the rectification has been done. And then you know, what is the sensor? Is the OCM sensor? Since these are given at every 15 days, they don't specifically mention the temporal resolution, etc. So this data is a safer motion set, which operates in eight bands. So eight bands near in the VNIR. Okay, so in the VNIR, there are eight bands with one kilometer spatial resolution every 15 days. Of the VNIR, NIR is taken and R is taken subtracted and divided by NIR plus R. Okay, so as I said, there is a 15 day window. So let's quickly view it. And then you could see if you view it, or we could see that there is some cloud cover, which is given in the white thing. And because in December time, we do have cloudy covers. If we, if you pick the year wise, and let's say 2021 June, approximately June view, there is less cover because in peak summer times, you don't have much cloud, thick clouds with with water vapor covering right condensation. So you have this, which is good. So I have viewed the June 2021. So you have that you can also zoom into a particular region. Okay, so we have zoomed in into Maharashtra again. And if you could see that minus one is the water bodies along the coast and wherever the water bodies are, there is minus around the water bodies still it's water reflecting. So it's red, but the dark greens and the greens give you the vegetation. So normally in June, there's not much vegetation, right? So we won't have much, but after June, after August, during the peak monsoon, so let's say September, first week, you will have more vegetation. So then there it is, you have now all these red areas are now covered with green. So you can also do this as a year wise product or individual products, which allows you to swipe and show you what swipe means. So let's go back to the same analysis. We pick June first week. So I'm going to pick June, and I'm going to say view. Once it views, then you can say activate swipe. Okay, we'll have to select the image. So the first image is there already June 15. So now I'm going to put, as I said, September, first week, because that is after the monsoon, there will be a lot of crops, right? June is kind of the peak summer. So or after the summer, June first week, maybe some rain will be there. Yes. So let's push it to May 31st, May 16 to 31st, get the view. See there's a lot of red color. Okay, we'll just keep it like that. And then now I'm going to take the second. So two dates we can take. So I'm going to take the second as September first week, activate swipe. So on this layer, what you see is the September 1 to 15 average of 2021. And if I move my mouse, so that's what I've activated. I've activated the swipe and deactivated the swipe. So if I move the mouse, nothing happens. I'm seeing the September month. But if I activate swipe and I move my mouse, then you will see that. Let's do it again. We'll have to go to land products, local coverage. Sometimes, as I said, it does get really stuck, which is good. I'm doing this again. So we'll go to May view. So we have this red color. And then I'm going to pick September first week, activate swipe. Now we have two images and I'm going to swipe. See, okay, there we go. So now if I move my mouse, you could see that it is changing. So if you could zoom into all these areas, all these red areas which have what does red mean? Red means not any vegetation and blue means water. So that is minus one, zero to 10 is really bad. So all these areas which had no crops. Now after the peak monsoon, you could see them turning green. So this is the monsoon irrigation, we can say. I've zoomed in and the process does take some time. It may take more time for you depending on the internet and the available you can see it says loading. I don't know where it is. So I'm just going to zoom out a little bit. Okay. So again, we have this going on, which is good. Now I'm going to take another time frame, which is the December first week. Why do I need December first week? That is the winter crop. So we can see how much the winter crop is done. Not much you could see, but it's good. So the first one is still being activated as a previous one. So I'll have to remove this and then view this again. So now we have the May month and then we have activating the December month activates swipe. So now if you go back and forth, the June month is the May month is there. The September month is gone. So underneath is a May end, which is the peak summer and then on top of it is my current December month. So December, which is the month, I could say the winter crops. So you could see that in the dual residue and with groundwater irrigation, there is some crops growing in this area in this central region of Maharashtra, UP, etc. There's a lot of green happening in this area. And that is all because of using of residual moisture and groundwater, especially if you could see here in the Punjab, Haryana region where there's a lot of groundwater pumping as we studied in the week nine end of lectures. So all this data can be helpful. So wherever there is an NDVI of above 0.5, whereas in the green color in this image, so this is a percentage for some reason, they have put it as full numbers. So it is divided by 100. So you can have this as all these areas where you have the red color, which is turning into green is winter crops. You can call it rubby in some regions or xythin in some regions, it's happening. So that is one. And then the second monsoon season that we also looked at is during March and April. So if we say that, okay, let's look at March and April, and then we activate the swipe. So now it has been loaded. So what you see at the background is for some reason, it's still the June data coming. So May is there. And then I'm going to look at the March data. Activate swipe and see you could see that going to pick May and May and View. So a lot of red, which means not much cropping going on. And then I'm going to choose March when I do it. So now you can see a lot of green. So these green, all these greens are where pumping happens, all these greens. And there is some monsoon also, but still you can see that all these red color gets more and more during May. So the summer season people normally don't grow much. It's too costly to bring water, pump water and put it. So that is why people refrain from taking this data, the cropping out. Okay. So that is how you could take NDVI. Again, NDVI would require you to download the images like we did in LULC classification. We downloaded the images and then pick the bands and subtract it and this is kind of an advanced level. So I don't want to go in depth of the process, whereas NDVI data is already available. So I'm just going to show you how useful it is. And you could see that quickly by these images you can do and you can download. So you can download this image or you can batch download and then do it. So I will have to log in and download and stuff, which we have already done in the previous classes. So I'll refrain from downloading it. So that is the one product. You can also look at the vegetation fractions and the vegetation fractions, how they are calculated, what are the equations, etc., is given here in the technical document. And you could go through and read it and also give citations for it. As I said, NDVI minus NDVI 0, NDVI infinity minus NDVI 0. And then let's look at the same month, which is May and then we can view. So it almost reflects the same, whereas the process is kind of an advanced NDVI. So you have some values more sharper and finer. But as I said, NDVI does the job, so you can always use NDVI. So this is good for understanding where the crops are growing in the winter period. And also, let me just do that again, activate swipe. So now we have the vegetation fraction of 100% in these areas along the against basin. And then some basins we don't have, right? So one kilometer resolution is pretty close still, given the fraction of farmers we have in the area we have. But still it works. So these are the two products in going. In the next classes, I'll go through each and every other resources that are available. Again, please go through the different programs. You will see the same product also given in different indicators, let's say terrestrial science. And then you'll also see the same here. The whatever you saw in the previous ones, the three OCMs, the four normalized, three normalized vegetation index NDVI and one vegetation fraction is also here. So there is a duplication within the website, but it's okay. It's just for you to look at and map it if needed. So NDVI is pretty important. It has been widely used. Please use it for understanding the winter crops area, the summer crops area. And if you need to find the area, what would you do? You would extract the pixels. So let's quickly show from this angle, do the same thing, but let's do it. And then let's pick September again. So each pixel is one kilometer, right? So if you pick a parcel within your village, if your village has a 10 kilometer area, a very rough 10 kilometer area, and four to three pixels. Let's say three pixels are dark green, which indicates that it is vegetation. So then what you do is, so if it is not going through a boon, just zoom out a little bit. It will bring it up. Okay. So the data gets populated. Sometimes it does take time. So let's remove this and then pick a new one. Sometimes if you want a new one, just go back and up and forth. Local coverage, pick a date, September 2021, and there you have. So as I said, let's say in the Gujarat region, you're doing it. And just zoom in and you can see, okay, this area, if you want to say, okay, how much of this area has been cultivated 100% very healthy vegetation, then you can use easily calculate. So one, two, three, four, five, six. So six pixels, six kilometers. So that is how approximately you could do this. But in visualization, I'm saying, but if you extract it into Google, your GIS platforms, especially the QGIS, then you can say that I want these to be clustered and then the clustering area can be removed. So that is a kind of an advanced level. I will not teach it for this class, but I'm just going to tell you that if you know the pixel area and the number of pixels that are in that particular class, you can easily multiply by the area of the pixel to get the total area. So with this, I will stop today's lecture. I'll conclude today's lecture and look forward for the next lecture on Google Earth Engine data for NDVI and NASA and Sentinel data for NDVI. Thank you.