Exploring Sentinel-1 for Coastal Monitoring

RUS Webinar: Coastal Erosion Monitoring with Sentinel-1 - COAS02

Estimated read time: 1:20

Summary

In the 43rd webinar of the RUS Copernicus project, Georgia Caradimo guides viewers on using Sentinel-1 data to monitor coastal erosion, focusing on Senegal's western coast. Coastal erosion is a pressing global issue affecting economies, populations, and infrastructures. The webinar explores the techniques for monitoring erosion using Sentinel-1 SAR data and the practical application of these techniques using the RUS virtual machine. Participants also learn the significance of using precise orbit files, thermal noise removal, and various data processing steps for accurate analysis.

Highlights

Georgia Caradimo introduced the webinar focus on Africa's coasts, particularly Senegal 🗺️.
Orbit file application and thermal noise removal are essential steps for precision 📈.
Participants can engage via the RUS platform to apply these techniques hands-on 🖐️.
Comparison of erosion with and without man-made barriers highlights significant differences 🌊.
Visualization using Google Earth enhances understanding of satellite data impacts 💡.

Key Takeaways

Sentinel-1 data is pivotal for understanding and monitoring coastal erosion 🛰️.
Coastal erosion is both a natural and anthropogenic issue affecting global populations 🏖️.
Utilizing SAR data requires specific pre-processing steps like orbit file application and thermal noise removal ⚙️.
The RUS platform facilitates user engagement with Sentinel data via a virtual computing environment 💻.
Practical exercises allow users to adapt methods to their own areas, making learning hands-on and versatile 🧩.

Overview

Georgia Caradimo welcomes the webinar participants and introduces the focus on monitoring shoreline changes in Senegal due to coastal erosion using Sentinel-1 data. Coastal erosion is highlighted as a significant global issue, impacting economies and populations.

The webinar structure includes an introduction to the RUS service, Sentinel-1 satellite operations, and a practical exercise using the RUS virtual machine. Participants follow along a thorough tutorial, emphasizing specific data processing steps necessary for accurate coastal monitoring.

Through the use of Google Earth, the session culminates in visualizing the impacts of man-made interventions on coastal conditions, driving home the practical applications of Sentinel-1 data for environmental management. Participants are encouraged to utilize RUS resources for continued learning and application in their own regions.

Chapters

00:00 - 03:00: Introduction to Coastal Erosion In the 43rd webinar of the RUS Copernicus project, hosted by Georgia Caradimo, the focus is on using Sentinel-1 data to monitor shoreline changes caused by coastal erosion. The session specifically examines the coastal erosion conditions in Senegal, highlighting the global economic impact of this issue.
03:00 - 04:30: RUS Service Overview The chapter discusses the impact of coastal erosion on global populations and economies. With 40% of the world's population living near coastlines, coastal activities contribute significantly to the gross domestic product of many countries. However, rapid erosion, whether from natural causes or human activities, affects the ability of populations to reside near coastlines and impacts the accompanying infrastructure. Monitoring coastal environments is highlighted as an important focus.
04:30 - 10:00: Sentinel-1 Data and Coastal Erosion Theory The chapter discusses how Sentinel-1 data and coastal erosion theory provide essential information for policy and decision-makers at local, regional, and national levels. It specifically mentions the northwest coast of Africa, where average coastal retreat rates of 1 to 2 meters per year have been recorded. The structure of the webinar is introduced, beginning with a quick introduction.
10:00 - 15:00: Exercise Instructions and Data Download The chapter titled 'Exercise Instructions and Data Download' provides an overview of a project related to roof service and discusses updates and satellite data editing. It includes an introduction to the study area and theoretical aspects of coastal erosion. The chapter concludes with a practical exercise using the ERUS virtual machine, followed by a Q&A session. The entire session, including the demonstration and Q&A, is expected to last between one to one and a half hours.
15:00 - 30:00: Processing Sentinel-1 Data in SNAP The chapter 'Processing Sentinel-1 Data in SNAP' focuses on an instructional webinar that is being recorded for later viewings on a YouTube channel and a training website. During the live session, participants are encouraged to send in their questions, which the team will address either during the session or will be discussed in a live Q&A session later. The emphasis is on ensuring all questions, especially those of high interest, are managed appropriately.
30:00 - 45:00: Visualizing and Exporting Data in QGIS and Google Earth The chapter discusses 'Visualizing and Exporting Data' in two key geospatial tools: QGIS and Google Earth.
45:00 - 70:00: Case Studies and Examples in Google Earth The chapter titled 'Case Studies and Examples in Google Earth' discusses various aspects of development activities supported by Google Earth's scalable platform. This platform is free, open, and equipped with powerful computing capabilities. It includes a suite of open-source toolboxes that come pre-installed on virtual machines, enabling users to manage and process data from Sentinel satellites efficiently. Additionally, the service offers specialized support through a user help desk dedicated to assisting with remote sensing activities involving Sentinel data. Furthermore, a comprehensive training program is provided to enhance users' proficiency in utilizing the data and tools available.
70:00 - 77:00: Conclusion and Further Resources This chapter covers the details of the training program and project updates. The services are offered in two main categories: one for users who have access to a Ruse virtual machine, and another for individual users and professionals working on R&D projects. Specifically, the chapter discusses various training activities available under these categories, emphasizing the role of Ruse virtual machines in facilitating these activities.

RUS Webinar: Coastal Erosion Monitoring with Sentinel-1 - COAS02 Transcription

00:00 - 00:30 hello everyone welcome to the 43rd webinar of the russ copernicus project my name is georgia caradimo and in this webinar we will learn how to monitor changes at the shorelines due to costal erosion using sentinel-1 data this time we will visit africa and more specifically we will check what are the erosion conditions at the coast of senegal coastal erosion has been a major problem for many areas around the world and the coasts are economically
00:30 - 01:00 important for many countries as a large part of the gross domestic product is derived from coastal activities and they are being eroded too fast about 40 percent of the world's population live near costs coastal erosion whether it is caused by natural or human activities it has important effects on the populations who can no longer live close to the coastline and to the infrastructure monitoring of coastal environments
01:00 - 01:30 and the tracking of the revolution provide fundamental information to policy and decision makers on local regional and national levels more specifically along the northwest coast of africa average rates of coastal retreat between 1 and 2 meters per year have been measured ok let's start by having a look at the structure of this webinar we will first have a quick introduction
01:30 - 02:00 about the roof service what the project is about followed by some updates we will talk a bit about sending one satellite edit data we will continue with a short introduction about the study area and some theory about the coastal erosion and then we will move to the exercise in the erus virtual machine finally a q a session will follow at the end the webinar will last around one to one half hour including both the demonstration part and the q a session
02:00 - 02:30 and please be aware that this webinar is being recorded and you will be able to repeat the exercise by yourself later once it will be uploaded in our youtube channel and our training website in the meantime you can start sending us your questions as soon as you have one and with my colleagues here we will answer to as many of them as possible the rest of them and those of high interest for everyone we will discuss them live at the q a session in any case all of your questions will
02:30 - 03:00 be answered after the webinar is completed in the relative document that we will upload at our training website together with this recording okay let's talk a bit about the rule service ruse stands for research and user support for sentinel core products it is an initiative funded by the european commission and managed by the european space agency with the objective to promote the uptake of copernicus sentinel data and support research and
03:00 - 03:30 development activities the service provides a free and open scalable platform in a powerful computing environment hosting a suit of opera source toolboxes pre-installed on virtual machines which allow you to handle and process the data derived from the sentinel satellites in addition to all that ruse also provides a specialized user help desk support to support your remote sensing activities with sentinel data and a dedicated training program let's
03:30 - 04:00 talk now a bit more in detail about the training program and the updates of this project we are offering our services in two main categories based on whether a ruse virtual machine is available to users or not the first category includes the rules training activities and the second includes requests coming from individual users and people working on r d projects regarding the training activities the rules virtual machines will be
04:00 - 04:30 provided to the users like this for the face-to-face trainings and the virtual classroom events we can accommodate around 20 participants providing a virtual machine to each one the type of the virtual machine depends on the training category and the ict team will support you during and after the session for any technical issues you may have after the event and in case you need to further use the virtual machine you can extend its total duration
04:30 - 05:00 maximum up to one month every month we are offering our webinars and the pdf tutorials will continue to be available to everyone via the roost portal for those who will receive a virtual machine for replaying a webinar the vm availability will not exceed the duration the two weeks for the external trainings requested 30 vms maximum will be provided per event and the participants will benefit from
05:00 - 05:30 the remote support of our ict team and now moving to the individual users that want to perform any project the standard ruse virtual machines provided until recently will not be available anymore but instead the solution of the docker container image will be offered it will contain the rus virtual environment and our ict team will provide the necessary support for us accessing the docker and installing the selected yours
05:30 - 06:00 tools on your own infrastructure you can find more information about the docker at the link provided here on the right here you can see there's websites where you can find all the information about the project in a groups portal here on the left so i recommend you to check it after the webinar to get familiar with the service and learn more about the updates of the project so register by creating an account and if you are eligible request the service
06:00 - 06:30 on the right part of the slide you can see and visit our training page where you can find the announcements for our upcoming webinars and events the recorded videos with the q a documents of the webinars delivered so far and our e-learning material there by simply creating an account you can access the material on various topics and practice at your own pace every time you repeat the course a different set of questions appear and
06:30 - 07:00 you have the chance to learn more at the end if you successfully complete the quizzes you will receive a certificate of completion in case you face any difficulty while signing up please contact us as i mentioned in the beginning we also have our youtube channel there's copernicus training where you can find all the videos of our previous webinars and you will also find this one in a couple of days now let's see some information about the
07:00 - 07:30 sentinel 1 data we will use in this exercise we will be using sar data provided by the sentinel 1a satellite the sentinel satellites are included in the space component of the copernicus program of the european commission and the european space agency the sentinel-1 mission is formed by a constellation of two twin satellites faced at 180 degrees to each other
07:30 - 08:00 it is an active sensor working on c-band it includes a right-looking active phased array antenna providing fast scanning in elevation and azimuth and it provides data with a short repeat cycle of six days and with different imaging modes here you can see the four different acquisition modes of widths we will use the iw and the four products of its mode of which we will use drd the ground rates
08:00 - 08:30 detected data we will now talk a bit about what causes coastal russia first of all the waves we have two types of waves the destructive and the constructive on the left you can see how the destructive waves work they basically have weak schwas and strong backwards and the strong backwards removes sediment from the beads these type of waves are steep and close
08:30 - 09:00 together on the right we have the constructive waves the characteristics of these waves are strong schwas and weak backwash with the strong wasps to bring sediments and build up the beads the backwash is not so strong to remove the sediment so the waves are also low and further apart in this slide you can see some examples
09:00 - 09:30 of various areas around the world as they are presented by jurian dick adult 2018 and the two areas on the left suffered from erosion while the other two areas on the right we have the position conditions also according to a jrc article published in nature's scientific reports it is stated that there is an increasing trend of erosion along the world's coastlines
09:30 - 10:00 with twice as much land lost than what was replaced over the last 30 years it is estimated that this covers a surface area of about 28 000 square kilometers so as we mentioned in the beginning we will visit senegal in africa and we will check how the coastline has been changing in the last six years using sentinel-1 data from 2015 until
10:00 - 10:30 2021 senegal is exposed to the atlantic ocean at its western part so we will set how these calls in different locations of the country are being affected we will move from northern coasts towards southwards and in order to be as consistent as possible to the weather conditions over the area we use data taken on july of
10:30 - 11:00 each year okay now let's move to the rose virtual machine in order to start our exercise so this is the rules desktop and here we insert our credentials and we login and this is how it looks like we have here our software that we need to use and before we start processing the data what is important is to download the
11:00 - 11:30 data so what we're going to do is we're going to go here in the firefox and we visit the copernicus open access hub where we will search the data we want to download we click here and we go here on the top right and we login with our account if we don't have an account first we sign up so here we have the account already so login and then we just navigate over
11:30 - 12:00 senegal and we plan to focus in this area that is exactly over here so this part i just select the drawing tool i draw the area that is of interest for me and then i go here on the left and i want to search the data available for sentinel one for each year
12:00 - 12:30 for the period of july but we will do it separately we will do it year by year for example let's set the parameters in the sensing period we go and we select a range for example 1st of july 2021 to let's say um 15th of july 2021 then we select the mission sentinel 1 and from this mission we want the product
12:30 - 13:00 type to be the grd and the central mode to be the iw as we said before and then we click search we see that this is the result of the search we run and in order to download this product we just click here on this icon that says download product and it starts automatically to download this product belongs on the 9th july 2021
13:00 - 13:30 once we have this we can go back to the search and go and do the same for the other product that we want of 2020 we set again the same period first of july 2020 to 15 of july 2020 and we wait to see based on these parameters we inserted what the search will return to us
13:30 - 14:00 okay now it returns two products what we would like to do is to try to keep the data that belong to more or less on the same time period so you can choose either image but i would suggest to take the image on 14th july 2020 what do we see now we see that these products are flying for those of you who are not aware what this is this means that any data that are older than one year by
14:00 - 14:30 the today's day or data that are not very usually used they go to the long-term archive they become offline so that they don't occupy a space and what you can do to make it online is to click here on the download product and once you click on that you get this message saying that the offline product retrieval initiated and what you need to do is follow these instructions and go and check after a
14:30 - 15:00 while in your cart if the products are online or not how do you find your cart you go here to your account and you click cart and you will find it in there note that you don't receive any notification when this becomes online so we just need to go after a few minutes and check if it has become online so we do exactly the same for the other products that belong to 2019 2018
15:00 - 15:30 2017 2016 and 2015. we cannot go back for sentinel one further back than that because there are no data available i will not do it now for all the products but in the tutorial i have the list of the products and if you want what else you can do is you can just go straight up here on the insert search criteria and you can copy and paste here the name of the product
15:30 - 16:00 and then search it by the name okay let me now close this because we do not need that i have the data downloaded so that we start processing the data in snap before we open snap i would like to show you the data we downloaded that we have copied them in the folder that we wish to use so if you open here the shared and you go on the training codes02 coastal erosion you will find in
16:00 - 16:30 there three folders one with some auxiliary data one that has the original data and the one that we will store all the data we will process if i open now the original one you see here that these are the seven images we have for our case study now i can open snap so that i load these products and we start working
16:30 - 17:00 okay let's close this message window and we go here in the original folder and we take them one by one from the oldest to the most recent we just drag it here on the left on the product explorer window and we drop it in there and it appears here let me load all the other products as well and then we will see each one of them where exactly they are located on earth
17:00 - 17:30 they should all be over the same area so we're here if we go here on the world view and we zoom in we can see that all our products are covering the area that is of interest for us which is here in senegal so if we go here and we expand this one and we also expand the bands folder we will open the intensity v8s
17:30 - 18:00 we double click on that and we create the image now note that this product it has intensity and amplitude bands for both polarizations the v8s and the vv from literature we found out that it's better for such cases to use the v8 polarization so i proceed like that and i open all the intensity vhs bands of all the products
18:00 - 18:30 so that we see exactly where they are located because as you understood we are not going to use the whole area this is not necessary so we need to see exactly how they look like so that we cut we subset only the area that is of interest for us now we have all intensity bands opened and the next thing we can do
18:30 - 19:00 is we can go up here on this window tile horizontally and then we can go down here on the navigation and select this option that says zoom all so we basically have the full picture of the area what can we see here we see that these five images the last five recent ones belong uh to the same area and cover the
19:00 - 19:30 exact same extent while these two others they are a bit different from 2015 and 2016. but which which is the area that we want to focus on let me go in this one so that we have a more clear picture the area we want belongs over here on this small area
19:30 - 20:00 in order to make it a bit easier i will just close a few windows so that we have some more space okay i think now we are okay so as i said in this area what we want to do is we want to zoom in a bit more and we want to create a subset we do not need the whole image
20:00 - 20:30 how do we create a subset there are several ways let me show you this one you go up here that says rectangle drawing tool you select it and you go over the area you want to subset and you just draw a rectangle once you draw it you go up here and you select the selection tool then you click on this area that you just drew and you right click on that and you select well-known text from
20:30 - 21:00 geometry here you have this polygon that has the geographic coordinates of these images what you do here you press ctrl and c so that you copy that because you need to save it and you go on the file here in the aux data and you create a file named expressions i have it already created for you
21:00 - 21:30 and what you do is you just go up here and you right click and paste it and you have the information of the polygon okay i have this done no need to copy it again let's go back and then you click ok once you save that you do not need this geometry anymore so you just click on it and you press delete and you delete it okay now i can close all these view images the view windows
21:30 - 22:00 and it is time now for us to go and start creating a chain of all the operators we want to use because all sentinel 1 drd data have some basic steps for pre-processing in order to be able to use them afterwards first part we do we click here on this icon which is the graph builder or we go to tools graph builder it's exactly the same
22:00 - 22:30 and this window opens let me just expand it a bit okay first thing we do when we have a graph builder we right click on the right operator and we delete it so that down here as you can see okay let me put it this way up here we have the read operator and down here you have the corresponding to that tab in there you insert all the parameters and everything needed
22:30 - 23:00 for the processing we will do it one by one we will add all the parameters we need in order to complete the processing but what we will not do now is we will not put any parameters down here on the tabs that will appear this will do it on the next step i will explain you why so first thing we do we need to add the apply orbit file operator so we right click on the white area you
23:00 - 23:30 go to add radar apply orbit file and then we select this arrow we drag it towards the apply orbit file and we connect the operator down here you can see the corresponding tab why do we do that this step is necessary because the orbit state vectors that are provided in the metadata of a sar
23:30 - 24:00 product are generally not that accurate and they can be refined with the precise orbit files which become available some days to maybe some weeks after the the generation of the product the orbit file now it provides accurate satellite position and velocity information and based on this information the orbit state vectors in the abstract metadata of the product are updated where is the abstract metadata
24:00 - 24:30 if you go here on the left you see the metadata okay next step we need to do we need to add the operator for thermal noise removal so let me add it and i will explain you what it does we go to add radar radiometric thermal noise removal okay so this operator why we use it
24:30 - 25:00 because there is this thermal noise in the tsar imagery which is the background energy that is generated by the receiver itself what it does it skews the other reflectivity to towards higher values and actually it hampers the precision of radar reflectivity estimates so for this kind of products in order to be able to use them we need to remove to remove this noise
25:00 - 25:30 once this step is done we need to calibrate the image what is the calibration let me add the operator by going here right click add radar radiometric and calibration and we connect this as well okay what is this why we need to calibrate the image
25:30 - 26:00 so for this kind of data the level one images we're using they do not include radiometric corrections and there are some significant radiometric bias this correction is necessary for the pixel values so that they can truly represent the radar backscatter of the reflecting surface and this way we can actually compare the different images acquired in
26:00 - 26:30 different times then then we want to apply a filter to this image so that we remove another noise let me just right click and go to add radar speckle filtering speckle filter and we connect it to the calibration so what is the speckle filter
26:30 - 27:00 the speckle is caused by some random constructive and deconstructive interference of the defaced but coherent return waves that are scattered by the elementary scatterers within each resolution cell so this reduction can be applied using this speckle filter and this is how we remove the noise of the speckles once we have complete all these steps
27:00 - 27:30 what we need to do is to terrain crank the data what do i mean with the terrain correction first let me right click and go to add radar geometric editing correction terrain correction i don't know if you remember but we saw the images how they look like before that i open them and we notice that they are actually upside down why is this happening this is happening
27:30 - 28:00 because the way the satellite is passing is always looking to the right and the first thing it says this is the first thing that depicts here when we open it so what we need to do in order to correct the data so that they don't look mirrored upside down north to south we need to apply the chatting correction also right now the data we have are in radar geometry
28:00 - 28:30 and there are always some topographic variations of a scene and also there is the tilt of the satellite sensor and sometimes the distances can be distorted in these images so in order to fix all these issues we need to apply the terrain correction and basically translate the radar geometry into geographic coordinates and having done that what we would like to do afterwards because what we care about is to see the
28:30 - 29:00 coastlines is we want to remove the sea from the image we want only the land to remain how we do that we right click we go to add raster masks lan c mask and in there we have the option to remove the c from the image last step we do once we have done all these parts we
29:00 - 29:30 just want to save the product we want it to be written so we go to add input output right and we add again the operator we had before so that this way we finalize our processing chain the other question is why we do this chain and we do not do these steps one by one because if i take the the first product and i apply the orbit file separately it will
29:30 - 30:00 store an output then i take the this output and i perform the thermal noise removal it stores a second output so if i do it one by one i have one two three four five six seven different outputs and this takes space this way it will run all the processors and it will store only one final output and we save actually both time and space
30:00 - 30:30 but as i said we will not provide any parameters down here in these tabs and this is happening because we want to apply this sequence for all the seven images we downloaded once you create this graph you can save it you can go here save and for example i save it here in our folder in the aux data and i have named it processing graph
30:30 - 31:00 and i click save having saved that you just click this window and then in order to apply this graph at once for all the seven different products we need to go to tools and click on batch processing so here we have this window let me make it a bit bigger okay first thing we do
31:00 - 31:30 we go down here and we deselect the option that says keep source product name because if you accidentally store the outputs in the same directory where you had the original data it will keep the same name and it will basically rewrite the original products you had so just to stay on the safe side deselect this option and then we load the graph we load the graph which is here and is named processing
31:30 - 32:00 graph okay we see that up here all these tabs have appeared now where are the data we want to process they are here on the left and we very easily add them here by clicking on add opened and this is opening all the products that are here on the product explorer perfect these are our products so now it is time to go step by step tab by tab
32:00 - 32:30 and set the parameters first of all we click on the subset we go here in the geographic coordinates and down here in this area we will insert the polygon that we saved before so if i go here to the expressions and i copy this part ctrl and c i go down here i press ctrl and v to paste it and then click update
32:30 - 33:00 then i can also click here and zoom sometimes zoom is not working so you can just do it with the mouse you can zoom in and you will see that there is this yellow rectangle that has been created over the area of interest that we said before and then we continue to the next step we go to the apply orbit file
33:00 - 33:30 over here make sure that you select this option of do not fail if your bit file is not found because as we said it takes some time for the orbit files to become available the updated ones so just to say to stay safe always select this option if we go now back to the subset tab now that we set this parameter in the orbit file we can see that as sourcebands we can see the amplitude the intensity of both polarizations
33:30 - 34:00 as we said now here you can see the whole product way better on red we have the original extent of the product and here on yellow is the very small area that we plan to use so this will save us a lot of processing time again as mentioned before we only need to work with the v8 bands so here we select the amplitude v8s we press ctrl and we select the
34:00 - 34:30 intensity vh2 we proceed with the next ones and here in the thermal voice removal we want to process only the v8 so we select only this one if you want to process both you don't select anything or you select them both then we continue to the calibration again you see now since in the previous step we selected only the v8s in the next step it takes only deviates now and this is the important stuff we need
34:30 - 35:00 to make sure is selected we need to have the output sigma not band selected this is what we need this will be the result of the calibration afterwards in order to remove remember this noise of this texture of this white and black we need to apply some filters again this is our band that will be created from the calibration and based on literature from all these
35:00 - 35:30 options we should keep this default one the least sigma here there are some other filters i will not go into details on what each filter does exactly so there are some slight differences but based on literature for such cases we know that the least sigma filter with window size 7x7 works the best there are more options it has to do with what information you have and how advanced you are so that
35:30 - 36:00 you select the appropriate filter and then change the rest parameters for this exercise we'll leave them as they are next we go to the terrain correction again in the source band we have only the sigma node v8s and here in the digital elevation model it's automatically the srtm three seconds we can go and select for example the srtm one second as you see also in the map projection
36:00 - 36:30 here we have automatically this option which is the geographic latitude and longitude this is amazing to keep it like that when we want to export the data afterwards for visualization in google earth but in this case since i want to insert them in qgis i will change the product the projection and i will use the utm wgs84 automatic and then once i click ok
36:30 - 37:00 it automatically finds the utm zone this area belongs i will show you in a bit later in case you want to export anything in google earth we can reproject the data afterwards only the final product we can show that later once you have set this information you go now to the lan c mask make sure the source band is the v8s
37:00 - 37:30 and you select here they mask out the c this is all you need to do and now you just go to write and as you can see here it is automatically keeping the original name of the product it adds the subset in the beginning and then at the end it is adding these suffixes orb for apply orbit file nr for thermal noise removal cal for calibration spk for spectral filter
37:30 - 38:00 tc for terrain correction and for last for land cmask it doesn't add anything but that's not a problem this is our final product now you only need to make sure that you have set the correct directory where you want to save the data in this case i have saved it as cos02 coastal erosion in the processing folder that we mentioned before now that all parameters are set you just click run
38:00 - 38:30 and then it starts automatically to run that graph for each of the seven products separately and once everything is done these products appear here on the left in the product explorer window i will not run it now but i will load the data that i processed already for you i can simply close these products actually i will select them all and then right click and close seven products
38:30 - 39:00 here we click no and now i go here on the processing folder and i will load all the products that have been created from this processing chain again i load them from the oldest to the most recent first 2015 then 16 17 18 19 20 and 21 now that i have them all here
39:00 - 39:30 if i expand them you see that we have much less folders than we had before and if i expand the bands folder we only have the sigma not v8s i just double click to visualize it and i do the same for all of them because i will tell you why we need this step to be done okay and this one
39:30 - 40:00 and that's it as you can see this area is way smaller than the one we had before again if we go here to windows tile horizontally we can see that based on the parameters we set the same area has been subset okay if you zoom in a bit more now you will see that some areas are black some are white and in general they go in
40:00 - 40:30 the gray scale if we now go up here to the pixel info and then we pass over some areas we will see that whenever we move the mouse on the left on the pixel info on the sigma not v8 here we're having a value so here you see the different values if we zoom in a bit more
40:30 - 41:00 over the areas that belong to coast and if we try to distinguish which area is land and which is water we will find if we move along and also based on some literature according to the values of the sigma naught we will find that there is a limit let's say at around 0.02 value of intensity of the sigma naught
41:00 - 41:30 that anything that is higher than that uh corresponding to water and anything less than that corresponds for sure to land this threshold might change so every time you need to check according to the area of your interest in this case we will keep the 0.02 for other cases it might need to be 0.01 or another number
41:30 - 42:00 so we go to the product explorer we close all these views we do not need them now that we see how this pixel info works okay and what we want to do is we want to mask out these values that are not useful for us for example let me do the first one let me open the sigma node of the product of
42:00 - 42:30 2015 and we will make a comparison first we right click on that and we go to band maths this is a very useful tool in the name of the band we can name it for example mask what you also need to do now is to deselect this option of virtual because if you keep it like that it will not store the data it will just show you the result so we deselect it because we
42:30 - 43:00 wanted to store the data and we click on edit expression and we write here an expression and we basically say if this sigma node v8 has values higher than 0.02 then consider it as no data value else consider it as the sigma node we
43:00 - 43:30 had in the beginning and you click ok and again you click ok and here we see that some pixels have been removed now if i put these windows close the one to the other you will see like for example we have the area in here we used to have all these pixels on the left the white pixels that on the right they have been removed
43:30 - 44:00 this is why we say just be careful with the threshold because if you use a higher one it will probably remove too much information that will be useful for you and you will lose the sense of reality now that you see that we understand that this corresponds way better to reality than the original sigma not without having been masked so very quickly now
44:00 - 44:30 let me close this one i will do the same for all the other products as you can see once you create the mask here on the left under the bands it appears so we will do this for all the products [Music] by writing exactly the same expression every time [Music]
44:30 - 45:00 okay and every time we do it let's go to window single every time we do that the mask appears here on the view window so i continue with the next product mask deselect this option edit the expression
45:00 - 45:30 note that if you do not do this step you will actually have wrong results and you will not be able to very well distinguish where the coastline is and we continue with the other one but always be careful in case you want to use a standard threshold there is not an answer that
45:30 - 46:00 there is one standard threshold you always need to check what is happening in your area and then select the appropriate one okay one more second oops i forgot to write
46:00 - 46:30 the criteria else this one okay and also note that every time that you run a bandmates [Music] you can see down here on the right once you write the expression like here if i just write if it says term expected i insert the term but it is missing
46:30 - 47:00 something afterwards so i continue with what i need to provide to it so that it understands and once you continue writing you will see that for example now that i have it complete it says ok no errors and you click ok otherwise it will not be working
47:00 - 47:30 and in case you're trying to find out why it is not working just make sure that you correctly read down here on the right what the indication is let's do this for the final product too [Music] okay and now that we have all the masks over here we are done with the pre-processing and we are ready to export these parts
47:30 - 48:00 as duratives so that we import them in qgis how do we do that the simplest way for example is let's go to the first one that corresponds to the 2015. if you just right click on the one that is opened here you go export view as image make sure that you select the full scene so that it takes the whole picture and then you go down here in the files
48:00 - 48:30 of type and you select geotiff tiff with geolocation and then we need to give it a name i gave it for example this one i gave it 2015 0 7 17 mask underscore mask and i did the same for the other products and then you click save make sure you save it in the correct path that you wish every time we repeat the same for the other six
48:30 - 49:00 products as well so that they are ready to be inserted in qgis and this is the step that i told you that you might want to export this information and put them in a google earth in order to do that if you right click and go export views google earth kmz you will get this error message saying product must be in geographic latitude longitude projection which was the default one but remember
49:00 - 49:30 we changed it to be able to use it in qgis so you click ok and you go to raster geometric operations reprojection and basically what do we do in here we take the images one by one all the seven you see that the reprojected is added afterwards you make sure you have the correct directory down here and then you go to the processing
49:30 - 50:00 parameters and you leave this geographic lat long as it is and you click run immediately this is reprojected it appears here on the left in the product explorer and then once you have this you just open it you open the band the mask band you right click on that like this and you export it as google earth games at and you store it okay i can now close snap so that we
50:00 - 50:30 pass the qgis i minimize this window and i open here the qgis okay in order to make it a bit faster i have already created a session in advance but what you have to do in order to insert the data is go here to the processing
50:30 - 51:00 and find these t files you have over here and just drag them and drop them here down on the left on the layers panel so that you load them together with that we will also use some auxiliary data we will use some data of the coastline as it has been approved and issued for senegal coastlines i will tell you also in the tutorial on where you can find these save files
51:00 - 51:30 so now in order to make it a bit quick i will just open the session i had created before so here as you can see i have loaded first in view mode the 2015 mask which is this one what i will do is i will put some background so that we have a reference and we can understand so we go to web open layers plugin
51:30 - 52:00 google maps google satellite and this i will put it at the very end so that it is not over my layers what we have done here for each one of these masks we have created a save file so that we have the coastline corresponding to this one so i will not play a lot with the zoom
52:00 - 52:30 because based on this google satellite background i have it is not loading correctly all the time but if i just deselect it and i zoom in you will see here that the save file we have it goes based on the boundaries that we have here from these pixels remember that if we had kept the data as we had them before before we mask them all this area would have been full of
52:30 - 53:00 pixels and it would have been giving totally wrong results if you want out of curiosity to export the unmasked duotif and put it here you will see the differences on the coastline and the results so how we create the cost line ideally in order to do that you need to have some validation with optical data with sentinel 2 or any other optical data you can have in this case we are processing only
53:00 - 53:30 the the radar data and afterwards we will insert them in google earth and we will see the comparisons over there and also as we can go only back to 2015 with this radar data we will exploit google earth and we will see the images that have been captured even 10 15 years before today so how to create this coastline
53:30 - 54:00 you go up here on the layer create layer new shape file layer in here you write the file name like i did here i wrote coastline 2015 and then you click on these three dots and you browse it where you want to save it in my case i wanted to save it here in the alps data folder that i have created in the course zero two coastal
54:00 - 54:30 erosion and you click save you see that the file name appears up here together with the name of the product you need then to set the geometry type from all these four options we want to select the line because we want to create a line we do not change any of the other parameters and we just click ok once we click ok it appears here on the left once it appears you go like i did now
54:30 - 55:00 and you click on it so that it is selected and it becomes blue and then what you do is you go up here in this pencil the toggle editing you click on that and now this layer is editable in order to create the line that should go across this boundary which remember based on the parameters we used in snap this is how we manage to distinguish to
55:00 - 55:30 distinguish the land from the sea and find this boundary you go to this option add line feature you click on that and you see now that from the hand it becomes this cross with a circle around so you then start you zoom in of course way more and you then start one by one like you see here and you go and you click following this
55:30 - 56:00 information you have on your raster okay i i will not do that i will cancel this one because i have done it already so once you finish and you have the coastline that you want you go again up here in the toggle editing and you save the edits a window will appear asking you if you want to save the edits so we did exactly the same for all the others for all the other rasters we have for
56:00 - 56:30 2016 too we have the coastline and then we proceed to the 2017 and we have the coastline and this is how we have actually all of the years going to one by one and seeing where is this area that distinguish the land from the sea based on the pre-processing we did in
56:30 - 57:00 snap okay now what i will do to make it easier for for you to see it is i will load the google satellite background but it's not working if i do not load a mask so i load it let me load the most recent the 2021 and what we have now is i will load all the coast lines that we have for created for all these years with each
57:00 - 57:30 one i set it to a different color and also like if you zoom here you will see different lines the coast lines and you can see that some of the areas are not that accurate that is happening because we're talking about sentinel 1 data it's not optical data and it is not an automatic process there are some commercial software
57:30 - 58:00 where they have some toolboxes available that you can automatically retrieve this information but this is not available here in what we demonstrate now the other thing that i want to add here is the coastline as it is provided from the experts and this coastline has been created if i'm not wrong in about 2012 this is a bit older so if you see i have here the coast of all senegal and you see that there are some slight
58:00 - 58:30 distortions sometimes okay it also takes a bit of time for the background image to load correctly so we might see some differences but what's what is important for us is if we go here and we check this line that we have the senegal coastline it more or less goes quite well with the lines that we have created the coastlines of the different dems
58:30 - 59:00 what we notice here which is not very clear over here that's why we will move to google earth is that some areas we have huge erosion while in some others we have the position and these beaches over here are sandy beaches so we will see why in some areas we have the position while in some other we have erosion as since the material is exactly
59:00 - 59:30 the same i showed you before how to export these raster these masks for using them in google earth but what we want to also export is also the coastlines we created here they are as save files and the save files they cannot be loaded in google earth we need to transform each one of them into a kml file so we go to the coastline 2015
59:30 - 60:00 we right click and we go export save features as and over here in the format on the top it will be selected the esri save file automatically but you need to go and from all this menu you need to select the kml the keyhole markup language again you set up a file name for example as we have it coastline 2015
60:00 - 60:30 you go here and you select the path you want to export it here in the aux data and you save it you leave the rest information as they are and you click ok and this is how you transform the save files into kml files to use them in girth so i can close qgis now and i will open google earth
60:30 - 61:00 okay how do we open the google earth in this ruse virtual machine we go here in the terminal and we write google does earth does pro and then we click enter and it opens the google earth in this case i will not do it this way i have already saved a session with all the information we want in google earth so i will close
61:00 - 61:30 this one and the terminal and i will go in the aux data folder that i have stored this information as kmz i have named it senegal kmz i just double click on that and it opens on google earth what information i have in here i have information of all the rasters as we did in qgis and i have all the coastlines
61:30 - 62:00 so here on the left let's see the options we have we have enabled the coastline of 2021 and also the mask of 2021. now if we zoom in we can see that our results correspond quite well to reality we face some small errors then we see some areas that have different results but don't worry we will go one by one to them and we will see some parts that i
62:00 - 62:30 would like you to know after you leave this webinar so why we chose this area this area is having a different reaction to the waves from area to area we can go and find another area that we see significant changes if we go along this one for example over here what do we have
62:30 - 63:00 here let's go to the today so this is the image we have today this area is important as you see there are quite some houses a lot of businesses and this is not the way it actually looks very natural let us go a bit on the past and let's start with what was happening in 2003 which is the oldest one we can go okay not very visible
63:00 - 63:30 2003 here you see that we see these structures this one this one over here let me deselect a bit this mask okay here we have this small area we can call it a mini port a small port and let's see how this area is behaving as the time is passing in 2003 we see that we start to have some um some sand here on this area on the beach but in 2009 you can see that we have way
63:30 - 64:00 more beats and this is happening because these man-made constructions are responsible for blocking the sediments so that they do not move and this way the sediments stay here and they create this beach also this port you see it starts to become operational if we continue after 2009 and we go to most recent years
64:00 - 64:30 we see that the deposition continues but you can see that but you can see that it depends also on the weather conditions that you have every time that sometimes we are losing a bit of sediment and sometimes we have some more that is being added so as the year passed we are in 2012 now you see that also in this area there is a lot of sediment that starts to remain here too
64:30 - 65:00 continuing even more you see how this changes from time to time and we see the evolution and if you want you can take older data radar data that are older than sentinel one and you can try to make this process yes as you see the more we come closer to the day the more sentiments we have on this area
65:00 - 65:30 deposited you also see that for some years we have more than one image acquired so if you want to process data you can easily take more than one image per year or even more than one per month and as we're racing close to today you see that this situation is now stable and it is safe for the area you see how these sediments are being
65:30 - 66:00 deposited here and this coast is actually expanding okay and we are almost close to the today you see that now we have even more [Music] sand if we now move to another area to use as an example let me zoom out a bit okay let's travel a bit along the beach
66:00 - 66:30 not this one if we go here for example you see this part we see how this area was and as we move closer to the today we will see that at some point in the past the erosion was so much that there were several and major issues in the area
66:30 - 67:00 so after we continue coming closer to today you see that some action has been taken like here again we have these man-made constructions and as we go closer to today we see that the authorities are continuing to expand this infrastructure so that we have the result we want and we start again to
67:00 - 67:30 create or even just maintain the coastline we had and here you see that this has been expanded because it was necessary to be done like that and after this has been created you see in here how much sediment can be accumulated also what is the advantage of that you can see that in here we have some boats some small boats
67:30 - 68:00 this makes the conditions in there even if it's wavy so that some boats can remain in the area and be safe and secure and this keeps expanding as you can see and the more these infrastructure are being created the more we have sediments being deposited on the beads as you can see sometimes we're having
68:00 - 68:30 some issues with this infrastructure either they are covered by water it also has to do with the timing the image was taken or they are destroyed and when we face such issues you see that again here all sediment is left and we are in 2017 where things seem to work quite well and we continue closer to today so that
68:30 - 69:00 we see how this area has evolved you see that in the past in here it used to have way more sediments but the last three years three four years we're having issues and we do not have these calls that we had before this is an area that keeps changing all the time as i don't want to waste a lot of your time i will just use a last example let me zoom out
69:00 - 69:30 okay so i'm just moving a bit uh southern to show you this example over here here we have this area where we basically see that the shoreline should be outside here but why is the shoreline that close to the buildings and we do not have it anymore and you can see how close it is to the
69:30 - 70:00 buildings and how much this erosion here is affecting the locals let us go a bit in the past so if we click here in the past and we go over here for example in 2003 we can see that the area used to have way more coast let's move closer to the day
70:00 - 70:30 2009 not that much of a difference and we continue 2010 we keep continuing 2011 2011 again 12. we see that we start losing a lot of cost in the last 10 years yes and we continue
70:30 - 71:00 13 14 and we start seeing the problem which is quite important in the area and we continue 15 almost there what do we start to see here here on this area we start to see these two infrastructures let's keep that in mind and let's
71:00 - 71:30 continue moving through the years 16. 17 still 17 we move to 18 we see that
71:30 - 72:00 the issue with the erosion is persistent and we continue we are only to 2018 now 19.
72:00 - 72:30 and what do we see now in 2019 okay we see this infrastructure so in 2019 after august these areas these barriers have been planted in the sea what will be the result of these let's continue to see
72:30 - 73:00 a bit later only one month later we see that the situation is not that bad anymore but if we continue second month afterwards and then we keep continuing three months later we see that some coast starts to be created over here and if we continue more this happened last year
73:00 - 73:30 as you can see you will see the evolution and as we go closer to today we are june 2020 and we continue what do we have down here let's see you see the difference here over this region it doesn't really have good results but over here in this region it has
73:30 - 74:00 amazing results let me just go a bit back you see july 2020 this is the coast here you see these barriers and how they actually break the waves and the result is if we move to the next image three months later four months later you see how well this construction is
74:00 - 74:30 working over here so we're having this beach that is being created the sediments are being deposited and we stop having the major erosion problem let's see this is one year past time from now if we continue we see that this area keeps keeping more and more sand on the beach what we also see now is that after
74:30 - 75:00 one year also this part it starts to finally work and this whole area is now prevented from erosion and as we go more closer to today you see that this is working that well so this sand has been accumulated and it is reaching this construction also why is that important because as
75:00 - 75:30 you can see if wisdom meaning that it's already people exploiting this area which is for touristic purposes and this is how we can prove that yes coastal erosion is happening yes there are areas that they are not eroding but we have sediments deposition and here are the ways that the local authorities can actually
75:30 - 76:00 use the technology use the knowledge we have and prevent the areas from erosion you can navigate for as long as you want um i think i will stop showing you now how this area looks like and i will close google earth and i will just need one minute in the presentation before we continue
76:00 - 76:30 let me tell you again that you can repeat this exercise by your own if you want to practice or adapt the methodology to your own application in order to repeat the exercise go to the roots portal register and apply for a virtual machine during that process you just need to specify the training code cos02 when needed you will receive the training kit that contains the pdf of the step-by-step guide and the auxiliary data
76:30 - 77:00 also if you simply want to download the pdf of the tutorial you can go to the roots portal and under the ruse library menu click on train with ruse select the ocean costs tab and you will find the cos02 tutorial uploaded just to remind you that all these auxiliary data mentioned with the coastlines the kml's and everything included in there expressions and the rest
77:00 - 77:30 they are available once you request eros virtual machine so this was all from my site here you can see where you can reach us for any questions you may have and i would like to thank you very much for joining