 So, our next speaker is Robert Nichols, good friend of mine on talking about a favorite topic of mine. Okay, thank you very much. It's a great pleasure and honor to be here presenting. And I want to talk really, I've been thinking about this issue of Delta, this couple of social democracy systems. I want to really take one project we've done. We've done a number of projects that have happened along this kind of issue. Two big ones, one's active. I'm going to look at the Espadeltas project, which is still ongoing in a way as I'll explain, but most of the work has sort of been done and it's done a lot of work about bringing together the natural and the human world in Bangladesh. So it sort of I think sets a nice context for the theme of this meeting. To give you a sort of plan, I want to talk a little bit about, you know, Delta's. Then look a little bit at some of the components we need to consider if we're going to talk about posted Bangladesh. Then the issue of integration. So this thing Delta DM will be mentioned a number of times. The Delta Dynamic Integrated Emulator Model, which is a framework for doing integration. I'll show you some results and then a few concluding remarks. And if we just start with Delta's, well, I mean, I think probably most of you would recognize Nile Delta without a title. And you can see the strong kind of contrast between the desert, the un-vegetated desert, and the very vegetated Delta. And it brings home the point that Delta's really concentrate ecosystem services. They're extraordinarily fertile places in the mid and low latitudes. And, you know, seven percent of the world population lives on them. And the key thing, I think, is that they can support more than 500 people per square kilometer. In Bangladesh, it's more than 1,000 people per square kilometer living off the land, tremendous sort of productivity. And this is a picture we've produced back in 2010 when we started this work, trying to illustrate the ecosystem services that are ongoing in the Ganges of our food for Delta. And, you know, they're diverse in terms of provisioning large amounts of food and water, et cetera, that can support the population. And the question, of course, is with Delta's under significant drivers of change, how will these evolve into the future? Then, ESPA, I've mentioned this ESPA-Delta project, I think it's important just to mention ESPA, so it's a kind of context. It was a 40 million, what it is, which is actually a 40 million pound program, looking at ecosystem services to poverty alleviation. So it's coming out to the millennium ecosystem assessment and looking at this relationship between ecosystem services and poverty alleviation or human well-being. And so it was very explicitly transdisciplinary. And our project, ESPA-Delta, formerly assessing health livelihoods and ecosystem services and poverty alleviation and populist Delta, was the largest of these grants supported, went from 2012 to 2016, and we worked in coastal Bangladesh. And we are actually continuing it at the present time with the government of Bangladesh in terms of applying it to look at some sort of public policy aspect. When a call came out, and it was sort of interesting, where in the coastal area is this facility? And it took about two seconds to think Delta and the rest is sort of the rest is history. Our study site, again you probably all know this pretty well, but we've got the Delta sitting here, the hatched area. We've got the catchments sitting above, draining the highest mountains on Earth. You must also remember the Bay of Bengal, because fisheries are very important in the study site. And we're actually not studying the whole Delta. What we're studying is coastal Bangladesh, which is this area here, including the Bangladesh portion of Sundabang, the Cullner and Barracel, an area with a population of 14 million people. Studying this was a very multi-disciplinary problem, and it took a lot of people. So this is just acknowledging the consortium that contributed to this project. There's seven UK partners, 12 Bangladeshi partners, two partners in India, and this is one of our consortium meetings in Bangladesh in 2014. And then we had this strategic partner which we developed for the project, which is the Planning Commission in the Government of Bangladesh, which was again built in, that we would engage with the Government and what developed was this strategic relationship with the Planning Commission over the last few years. What was the idea? We had a sort of vision from the very beginning, it was very integrative, and we wanted to really allow policy makers to sort of use information over this large spatial scale to think about ecosystem services and people's livelihoods. You might say in Spain English, link science to policy at the land-state scale. It's sort of at this very large spatial domain. And how would it make a difference to the poor? The money for ESPA, I should have said, is coming largely from different, essentially like USAID. So it's very much linked to development. So our pathway to impact, which you have to have when you're doing development, was to develop methods that would inform the national government so they could make better decisions across this land-state. So that was our pathway to impact. When you start to think about the different scales that are going on, I've already hinted at some of the different scales that are there with the map, but we've got to think about global scales, things like obviously climate change, Bangladesh is always the host of child for the impact of sea level rise. But we must also consider regional and river delta scales. There's lots of changes, such as dams. Also DACA, I mean DACA was 250,000 people in 1950, it's now 20 million people. It's not in our study area, but it's having a huge effect on the country and on the delta. So it gets quite complicated when you start to think through all the different things that are going on. So in terms of our approach, we wanted to look at both, we wanted to understand the present day situation and then prognosis, be able to think a little bit about the future of these issues that we're concerned about. So first of all we had to think very much about the integration of social, physical and ecological dynamics and think about what mechanisms are actually causing them, so a sort of systems view. And then so in the project we actually did a lot of participatory work that was quite important in terms of engaging throughout the project with the Bangladeshi stakeholders in government. So we held workshops at least once a year, maybe more frequently than that. We're going to say a bit more about that in a moment. Then we had to look at what were the issues of physical and biological processes that really benefit people living in the delta and the rural environments. Actually think about quantifying these relationships and then bring it all together into an integrated framework and model to be able to explore possible futures. We managed to do all those things. In a kind of big oral sense, this picture here captures how we designed the project and we have this aspect of governance analysis and stakeholder engagement at the top. It's long and thin because it ran through the project. And we started with conceptual analysis because in many ways the starting point we weren't really sure about the relationship between physical services and human well-being so we did some qualitative work. This identified that seasons are very important so it changed our methodology. We were going to do one household survey, we did three for the seasons to capture that dynamic. Then that informed the social and economic data collection for the household survey. Also a lot of biophysical modeling, I'll explain that in the next slide, which moved on into an integration phase where we developed scenarios and we also developed an integrated model. I'll explain that more in a moment. That was delta DM and then this idea of policy analysis where we simulated the futures with that model and we went back to the stakeholders and so the stakeholders were helping us develop the scenarios so that they were very much buy-in with developing the scenarios and then taking results and actually discussing these results with the stakeholders and actually what would they do? If they didn't like what they saw, what would they do? That's the best idea of an iterative learning which I'll mention again. So what are the components that we considered? Well, there's sort of under three different areas. We looked at governance analysis and the stakeholder engagement and that's about ongoing process. We looked at social economic data collection and analysis. This included the household survey. We did analysis to the census, to a specific model we built from that. We looked at population projections. We looked at economic trends over the past decade or two. And then we did a lot of biophysical work as well. The climate, and this includes sea level, both temperature and precipitation, CO2 as well, which is the agriculture modeling. That then fed into the upstream basin modeling and each one of these boxes here represents essentially a model or a set of models that were coupled and then a Bay of Bengal analysis. And then this feeds down into analysis at the scale of the delta. I'm looking at things like salinity, morphology, land use, and then into provisioning ecosystem services, agriculture, aquaculture, mangroves and the fisheries. And so this was all done in a sort of loosely coupled way but then to really see through our vision of prognosis we needed to actually think about how we could join and link all these things together. Because it's very ambitious just to do those three elements. But then we wanted to have the fourth element of actually joining it together into a truly integrated framework which is the Delta DM. And I always say is Delta DM, is this a piece of software that sort of describes coastal Bangladesh or was actually a social device to get people with very different views to actually sit down and talk about their understanding and develop a shared understanding. And of course it's both because that's how it was produced. But it spends an awful lot of time and energy in doing that. So if we look at Delta DM in a little more detail it's an interdisciplinary tool. It sort of covers a lot of different domains. It builds on the high fidelity models that have been used in the sort of in the biophysical work, things like Delta 3D, EpiCom. But it's also taking secondary data. It's using the household survey and it's taking the expert knowledge of people within the consortium and the stakeholders. What it actually corresponds to is it's really a metamodel. It's a hybrid. It has some aspects of it are lookup tables. Some aspects are statistical emulators which is why the word emulator appears in the title and in other cases we're actually running the code, for example, the agricultural model. But it gives a fairly quick running time. It harmonizes scales and methods. It's fully coupled and it gives, we have the feedbacks and it's giving us a quick running time. Who's actually participating in this? Again, the people involved is I think it's very important in this kind of process. We've got our stakeholders, 55 agencies. We've got the household surveyed. We've got the specialist team with a wide range of different skills. We've got the integrators. These are a subset of the Delta team but they're the people that are actually thinking about the integration. And the Tilla-Lasares in the room has actually led the integration in the project. And then lastly, we have the users. And they're sort of taking it in terms of applying the model and taking it back to the stakeholders. And thinking about how these sort of link together, when we start building the model, we think about this long iteration route that sort of involves advice from the broad team and the knowledge that we've developed. But once we get a model that we feel represents the processes that are going on for the questions we're asking, we can then think about a much shorter iteration where we don't need to go to the expert teams and we can just run Delta DM. And then we have to keep on asking the question, are we using the model within the domain where it's valid? And we have to remember we might need to go back that long iteration if we start to ask questions that are outside the box that we've actually sort of played with up until that time. It takes a lot of different inputs in Delta DM and I won't be able to read through all these, but you can just see them. But there's a wide range of information, both climate, various biophysical dimensions, the ecosystem services, the demography, et cetera. So taking both natural and social science inputs to be able to run the model. And these are expressed from the outputs of the more complex models or the emulator, et cetera. And so we need scenarios to do that. And then the main outputs, when we actually output the hydrology is obviously fundamental to the sort of ecosystem services in the Delta. Stabilization, salt is fundamental to agricultural productivity, issues around livelihoods. And then an unusual, probably an unusual set of outputs that normally aren't output in the model in the sense that you'd stop there and you'd hand your results over to another domain expert and they would do an interpretation. Well, this is all output within the model. So we have a lot of wellbeing and poverty and health indicators both at the scale of the household and obviously, you know, you can, if you have a household output you can then aggregate up to regional outputs. So things like income, relative welfare, et cetera. Things like genie coefficient, which is the inequality of income within a region. So aspects that were done within the model, I mean, we did things like the sort of, we brought in biophysical emulation of these types of aspects, primate hydrology, mangroves, et cetera. And then certain aspects were developed within the model including like a salt balance, prop watt was extended, and the emulation of the numerical models. The household component was built on the empirics of the household survey and it's an sort of agent-based model where it considers 30 plus household archetypes. There's only a small number of things people can do in Bangladesh with about seven or eight different livelihoods but they can change what they do between seasons. So you get a large, a large range of possibility space. It looks at economic decisions and the poverty health indicators and outputs. Obviously verification and validation are sort of critical and look into the bugs and also checking against other data sets. And here are sort of some examples of validation. This is sort of comparing EpiCom with Delta DM for salinity, that's for surface water salinity for soil salinity and then for sort of crop productions I'm sort of looking through these very fast but just giving you a flavor here. And equally for the sort of social, some of the more social indicators, the black line shows the mean output from Delta DM and the rays shows the spread and then the circles are showing different measures independent from various surveys within Bangladesh. So we're looking at, here we're looking at sort of, make sure I get the total expenditure in the family in A or down here, we're looking at the Gini coefficient. So again, reasonable fits to observations. And note that this is people living on less than $1.5 million per day so the numbers are falling both in the model and observed so people are getting wealthy in the Delta, that's important. Our scenario framework was based on a sort of development scenarios which were developed in a participatory way with stakeholders and then climate scenarios and we had a sort of 3x3 matrix and we had this idea of an iterative learning loop whereby we start off with experts developing knowledge stakeholders developing the sort of scenarios at the Delta scale, then a quantitative translation of the information that was provided by the stakeholders because often they're providing narratives really about where they are and where they might be in the future. This is then fed into Delta DM, reducing output, taking that back in a form that can be understood by the stakeholders and then there's a sort of cycle of co-learning with the stakeholders. And the three scenarios I'm going to show you are that sort of access there from the least sustainable and the highest climate change up to the most sustainable and the least climate change although they're all similar amounts of climate change. And two observations I just put in listening to some of the earlier talks that we often wanted to really engage with our stakeholders but the workshop can only be one day and so you've really got to spend your time very effectively and plan how you can do that and the narratives are a sort of key communication wise that word has been mentioned in other talks but the narratives I think are very important and that's how a lot of this exchange of information and knowledge go take place. Scenario examples, again a wide range of scenarios are sort of fed into the Delta DM model just in the summaries here and I'll just illustrate a few of them here. So again fisheries, this is an output of a model by Phyllis Marine Labs and it's sort of showing that fisheries are likely to fall but the difference in the future, this is out to 2050 is largely a function of how you manage the fishery. If you manage the fishery in a sustainable way you can sustain it although there will be some decline even if you follow good management practices. Demography, well population actually you expect population to be rising population is either slowly falling or maybe falling quite a bit so that this area is losing people without migration going on and economic scenarios, well there's much more growth in the non ecosystem service based aspects that red line is showing that the we expect based on observed trends to see more percentage change in economic output in the non ecosystem service part of the economy. Just to show you a few illustrative results very very quickly, this is flooding in 2050 at the end of the monsoon the polders were shown in some of the earlier talks and so the areas here are flooding are outside the polders, very very similar for all three scenarios and then we can actually even play around with the polder hikes, we can look at the dikes we can raise them or we can lower them raising them makes no difference so it's suggesting they're high enough you're not going to get any benefit at least with these water levels this isn't a particularly extreme year if you reduce them clearly the flooding greatly increases and then obviously maybe that's what you do with tidal river management but in a rather controlled way this is just removing the crop yields this is sort of aggregated crop yields the top line is showing over the whole year the bottom is showing in the dry season and so from the least sustainable to the more sustainable you can see the colors are getting less red and orange we're seeing higher productivity there with provisioning ecosystem services we can aggregate the information and you can see that again the least sustainable to the more sustainable this is showing social economy again it's higher in this side income is higher on this side and then with the hazard based ones a lot more variability maybe to expect from sort of wet to dry years etc storms more variability a less maybe less clear pattern there the livelihoods can also be unpicked and these are just sort of showing six out of the more than 30 archetypes which is showing different components things like remittance is yellow business is orange these are just different types of ways people make their livelihoods and ecosystem services are becoming less important in time and you can actually even look at there's a multi-dimensional poverty index which Attila and Helen Adams have in review at the moment which shows different people are always poor people who are getting richer people that are going in and out of poverty and people that stay poor you can sort of see different typologies of people within the system so just sort of tie it then and how does this make the policy well we have this stakeholder engagement and so this is one of the events we held in the planning commission in Bangladesh and as we were doing the project this Bangladesh Delta Plan came along which is funded by the government of the Netherlands is taking a holistic much more holistic view of the future of Bangladesh than has been done before and our project they made the comment well actually before this nobody's really thought quite so holistically about Bangladesh as you are in the S for Delta project so we sort of naturally found strong engagement and they facilitated a lot of our workshop which really helped with stakeholder engagement and there was even a final event which was looking at the Delta Plan and our work and they actually from that have written to DFID which is very supporting us and we are now actually starting a new project which is going to be looking at applying this to assess selected projects within the Delta Plan and it's a proof of concept so using it in a policy manner rather than a research manager assess multiple indicators of these projects including things like poverty things that maybe wouldn't be assessed typically in a project appraisal in terms of the kinds of things we can assess it can be engineering but it can also be subsidies new channels lots of different things can be assessed within the project and within a small project we are doing a side project called Reach we are looking at very different strategic policies for flood management in Bangladesh raising polders tidal river management we heard about earlier or even retreat just looking again, what if kind of scenario so conclusion, last slide this work I think provides a new link model and data framework for thinking about the future of coastal Bangladesh and the Bangladeshis have commented that this is something they haven't seen before and the people we are working with have engaged very strongly with the policy process as a result they weren't working before we started working I think it has a modular approach that allows incremental improvements so it can progressively change and be improved working with stakeholders was fundamental to the success of the project so far I think we had regular workshops people saw that we were taking note of what they said, it had local ownership they felt it was their analysis not our analysis not us as experts coming in and just imposing our view the issue there's clearly things, I'm very pleased we've got this but the issue of understanding uncertainty and sensitivity of the coupled model as an ongoing process all the individual components we understand but when you start putting together clearly there are questions there and we're learning as we go in terms of that three kind of key take home messages what have we learnt from all this huge effort well in the future up to 2050 more influenced common choice policy interventions and climate change but Bangladesh tends to get written off people say why should we invest there should it be gone, let's not bother well I mean there's a lot you can do in the next 30 years you can build a very different and wealthier Bangladesh which would be much less fundamental to the climate change but ecosystem services diminishes a portion of the economy with time that's continuing historic trends we try to really hit that and it doesn't go away it seems to be a robust result and significant poverty persists in some locations under all scenarios the development is still going to be important the natural economic growth you're seeing in Bangladesh is not going to get the problem to go away and I suppose like last thought, I think engaging with the kind of policy questions I've shown you here to my mind is a very good way of forcing consideration of the human dimension if you have to engage if you're going to think about the sort of policy questions you have to engage with human dimensions in some way which depends on the question you're posing thank you very much I was hoping that we'd end a conference set of talks on the dynamic dual and you did it fabulous so questions yeah Bob it was a fascinating talk I guess I have one question coming from the human dimension side and the way that the scenarios were used I'd be interested in learning a little bit more and I know there's only so much you can say now but you had exogenous assumptions about population and overall economic growth did the models how did they then translate that like I noticed your $1.90 a day statistic and I'm just curious were those things that were put in the scenarios or were those things somehow calculated in the models and any indication of how the demography the demography and sort of some of the economic growth scenarios they were scenarios that they had to be I guess with this kind of problem you always have to ask where do you draw your boundaries I mean it would be lovely to have a model where where the demography was an emergent property of the model and actually yesterday Attila Lassabra sent the poster on another project called DECMA where actually we're looking at human migration and there we're actually also going to have an economic model so maybe in the future we might, you know, again it's not planning through at the moment but if we can fuse these two together we can have that demography and the economics built into it but they were scenarios in this analysis because that was all because of resources other questions I'd like to ask also as your as your scenarios are kind of projecting towards increased independence from vulnerable ecosystem services and energy as a key part of that how explicitly did you model the transition of the energy infrastructure or was that just implicit in your economic models that was implicit I mean yeah we didn't mean I mean that would again that's another thing that we're working looking to the future it would be nice to see a merging of those to sort of bring those in one of the actually in the context of Bangladesh one of the biggest aspects of the low hanging infrastructure these bridges and transport network because the work on the sensors demonstrated that access was a big factor driving poverty along the Padma so building, so actually just improving those transport networks again it would be a fairly simple in fact they are building bridges obviously everywhere but I mean this would strongly support that so we have a break now after