 So, my name is Louise Gloss. I'm a professor in the Plant and Microbial Biology Department at UC Berkeley. And it's my pleasure to introduce this session on tools, technology, and research techniques, bottom up approaches. And our speakers are Nico France from Arizona State University, Jesus Pinto Ledzima from University of Minnesota, Christine Wilkinson from the University of California, Berkeley, and Elise Zipkin from Michigan State University. As all the other sessions, we will hold question and answers tell the all the presentations are finished. And so it's my pleasure to introduce the first speaker, Nico France. So, Nico is a Virginia M. Oldman professor of ecology and bio collections director at the School of Life Sciences at Arizona State University, and principal I'm the investigator of the National Ecological Observatory Network, neon biorepository, and the I dig bio, I love that name, symbiota support hub. Dr. France is a evolutionary biologist and insect systematist with the research program focusing on developing innovative bio collections infrastructure, and bio, biodiversity data sciences services. So thank you very much for joining us and take it away. Thank you so much, Louise. Let's see how quickly I get here. Hi everybody. Well, thank you for this opportunity. Can you see my screen okay. Yes, we can. All right. Great. So welcome. I think this panel and this session on the bottom up is really featuring a biodiversity data, biodiversity collections and bottom up approaches. In that context, I hope the following 10 minute presentation on the symbiota software project and how we apply it for example to the neon National Ecological Observatory Network, biorepository data portal will be informative. So just a quick overview, there'll be a little bit about what the symbiota software platform is, how it integrates with other organizations than a more concrete implementation in the biorepository portal, and hopefully a little bit of time at the end to talk about new services. Maybe to get that out of the way, I am fortunate to be part of two intersecting teams at Arizona State University Biodiversity Knowledge Integration Center, what we call Biocape, both the neon biorepository team and the symbiota support hub team. These are both separate but connected NSF funded project and the teams are greatly acknowledged. What is symbiota? First of all, it is an open source software system. You can see the URL here for the open source code right there that interacts both as a content management system for biodiversity collections, as well as a mid-level aggregator. So the bottom up part is definitely true here because it starts with individual collections that come together in themed so-called portals. These are self-identified or self-identifying communities of practice that contain data from multiple sources. There's typically multiple different institutions that come around along a taxonomic or geographic or ecological or otherwise theme. These portals are community managed, so they are grassroots bottom up efforts and the portals themselves can either contain a particular collection as a snapshot, meaning that it is static or can be periodically updated, or as a live managed system, meaning that for that particular national history or ecological collection, it is the day-to-day primary management system. And so the content management system in and of itself is relatively light in terms of the IT infrastructure support that a participating collection has to have. It has a web-based data editing interface, a suite of data cleaning and curation tools, and supports multiple remote users directly through a browser. Here are some of the 55 and counting and growing symbiote portals that have so far come into existence and which sustain more than 1000 live managing natural history collections, which we believe at this point is more than 50% of the market share in North America at a minimum and currently rapidly expanding into Latin America and other continents. So as you can see here, these portals can have a vertebrate theme or they can have certain regional themes and so on and so forth. There's a paleo data support with at least one or two portals that's being added gradually. Here are just some of the tools and the features of the symbiote platform as illustrated through snapshots from various portals. So they're extensive public search and mapping support for what are called occurrences, Dovincorps standard records. There are features to curate an author, basically checklist, species checklist that have a regional and taxonomic constraint. This is a quick snapshot of what the data entry form looks like for those inside a collection or with the proper access rights that are editing this. So there are taxonomy support tools. There are image annotation tools and there are standard compliant other data entry tools as well as opportunities to link to external resources, link to images, specify trades, update taxonomy and so forth. There are also built in geo referencing tools and not to get too much into detail because time here is limited, but symbiota.org is our primary URL is actually not on the slide but you will find it as soon as you put in that name. Now the symbiota support hub is different in principle from the open source software because it is actually a five year NSF supported sustaining grant in the DBI division of biological infrastructure and the symbiota support is part of IDIK bio, which is a national hub for digitizing collection that's at this point in its 12th year of existence. And so as part of this symbiota support hub and part of IDIK bio, we have created fairly comprehensive community support services and have had the opportunity for example to to run a highly active help desk, help desk, help desk with I believe more than 3500 tickets just in the last 15 months. So a lot of community engagement help and tutorial items and also discussion in Q&A board. So if you want to learn more about that, just go to symbiota.org and that's all the general sort of introduction that I will have here. Now on to an implementation of one symbiota portal specifically for the neon biorepository. And so we recently made these nice sort of postcards that your R code should actually work if you have time to use it and it will take you directly to the portal but again, if you went to biorepo-neonscience.org you would also find it. So neon very briefly is a continental scale program to monitor and forecast ecological change. Neon uses standardized organizational and environmental data and sample collection. There are 81 sites that have been running since 2019. So all the way up to 2049, including terrestrial and freshwater sites within 20 ecoclimatic domains including Alaska, Hawaii and Puerto Rico with a steady consistent sampling regime. Here's just a brief map. Many of you may have seen this before of how neon divvies up these regions and basically monitors at the 81 field sites. So then the biorepository which came into existence in the fall of 2018 so has been running for about five years now has a very defined scope of work which includes to receive all of the samples at each neon site up to 182 data products are being generated and about 65 to 70 of those actually have a physical sample. And these samples come to the biorepository at ASU which is a physical facility with about 27 and a half thousand square feet currently. So our mission is to curate roughly 100 case samples per year, maintain the data integrity, curate them according to best practices, publish the sample associated data and our semi portal and most importantly make them available for third parties who have specific ecological and evolutionary research questions oftentimes with additional funding in order to leverage these samples for that research. So then this is a complicated figure, but very briefly from our perspective of operating the biorepository, the researchers are in the center. And the researchers will be making requests to our biorepository sometimes through our biorepository data portal in order to receive samples for their respective studies so we put out loans for this community. Whenever these requests come and they get samples, they also have to get data associated with them, which means that our physical and our virtual environment have to be of course very, very closely integrated. But where do the samples come from? Nico, there's one more minute. Okay, only one more minute. That's terrible. Okay, I will leave this here just to say that involves a lot of coordinated software integration and you can see that there's these application programming interfaces that are very critical, but they're only the technical part. The social contracts that go with them are really essential. This is how many samples we have received so far, according to the different categories. This is an overview of the number of projects that we have so far supported but we are capable of supporting possibly twice that volume per year at this point and we're proactively looking for that. And the last part that I have here is about the Doveringport standard. I do not know if that came up already during today's discussion but this is a very, very prevalent standard in the biodiversity collections community. And one of the features that I will rush through that we have built for Neon is to basically translate ecological data that we get from the domain facilities through a number of harvest and expansion and production calls into Darwin core records, which then look like this these are so called extended specimens that have globally unique identifiers, lots of associated metadata and linkages out and enrich. Therefore, each of these occurrences and also making accessible, given that this is an ecological project to a much larger, larger biodiversity community. They also receive citations for specimen usage. They also receive citations for taxonomic identifications, and we recently built an outlet to link these data to a protocol the environmental data initiative. And with that, I will stop. Thank you very much. We should go on. That was terrific. So we should go on to the next speaker. The next speaker is Jesus Pinto Ledzima from the University of Minnesota, and Dr. Pinto Ledzima is an evolutionary and quantitative ecologist whose work focuses on developing a deeper understanding of species coexistence and patterns of diversity across terrestrial scales and the underlying processes that drive maintain and alter these patterns. He's a postdoc role associate in the Department of Ecology, Evolution and Behavior at the University of Minnesota, whereas work focuses on terrestrial systems. So thank you very much for joining us and we look forward to your presentation. Hi. I would like to see my presentation. Yes, perfect. So I thank you for the invitation for the kind introduction. Today I'm going to talk about how we can integrate multiple dimensions of plant biodiversity. And I want to start this presentation with this slide that so we know that our environment is changing rapidly. And this following here is from a paper room of Brema Gill published in 2015 that show how human activities is changing the landscape and affecting the species that are in there. So, but also besides the land use change, the climate is changing really fast. For example, just in the last July was the hottest July in more than 100 years. So the combination of these factors are affecting negatively the distribution of the species but as well the abundance. In fact, thousands of pieces are under threat of extinction as of today. And some researchers even have highlighted that we probably are living in an era of the sixth mass extinction. So in order to understand these changes on biodiversity and to allow us to monitor those changes, we need to find new ways that allow us to characterize and monitor biodiversity. So remote sensing technologies and products represent a good alternative to do that because it capture information from the earth at various temporal alespatial scales. So in this presentation, I'm going to give two examples in which how we can integrate remote sensing products with in situ observations. So to start, I want to show this map. So this map is simply the distribution of weather stations across the world. In here, this information from the weather station are used for creating these climatic interpolated surface surfaces that ecologists used to predict the distribution of the species and as well to predict the changes on biodiversity. Noticing here that the sparseness of the data, so on this sparseness is mostly located in tropical areas where it's usually the places where more species are there. So besides of that, this sparseness in the data can produce uncertainties in our predictions. So in order to reduce these uncertainties, so KT in 2015 and collaborator proposed an approach in which we can use remote sensing products to predict the distribution of species, but in this case, the use of the remote sensing technologies would allow the more accurate prediction of the species composition and the distribution. So taking advantage of that with my collaborator, close collaborator, Jenny Cabender-Vares, we use data from the sensor modis and we created a new set of environmental variables. In this case, for the continental United States, and we call this environmental variables Earth observation bio-climatic variables. This fully satellite or remote sensing variables resemble those bio-climatic variables as those are similar for the off work link. This information then was combined with occurrence data to predict the distribution of ox species in the continental United States. Here are part of our results. On the left hand, we can see the different variables that predict the distribution of ox species. And on the right hand, we can see that if we stack all the distribution of species, we can obtain a map, it's an ice map that showed the number of occurrence species per grid cell. And the bottom is the uncertainty in our model predictions, not in here that the areas that have more ox species also present high uncertainty. But we didn't stop just there and obtaining good maps or nice looking maps. We wanted to explore if our predictions can resemble the observed data. So to do that, we use the data from the National Ecological Observatory Network and test our predictions at two different spatial scales. One on the scale of the macro-ecological scale or the scale of site and the ecological scale or the scale of plot. For our on the x-axis, we can see our predictions using the space distribution models and on the y-axis, the observed data. At a macro-ecological scale, we found that the predicted and observed data are highly correlated. However, at an ecological scale, this association, although still positive, it decreased. So is there a scale dependence in here? It is probably so because the biological processes that generate and maintain the patterns of biodiversity change with the spatial and temporal scales. In the same vein, the detection and interpretation of the patterns from using satellite information will change with the spatial extent and pixel resolution. So to try to solve this problem of scale, instead of using satellite information, we use airborne data as well from the National Ecological Observatory Network and develop a framework that allows us to estimate similar metrics of biodiversity as those from metrics that are used in ecological studies. Our results indicate that, yes, in fact, we can have a high correlation in between the metrics that are derived from remote sensing, airborne remote sensing, and the metrics that are observed using observed data or the ground. In this case, those are metrics for alpha diversity for the taxonomic, droid diversity and phylogenetic dimensions. So the next question here is, can we use these models to obtain predictions about the diversity? The short answer is yes. Here we can see the observed versus the modelled diversity metrics using airborne data. So in order to better quantify the spatial variation of the biodiversity, we need to find reliable approaches that allow us to detect those patterns and appropriate scale. So in fact, oh, sorry, scale matters. So although I'm presenting here a series of challenges and recommendations, I want to focus on these last two. So if we want to integrate or recover the information on the ground, remote sensing itself will not recover the processes and the mechanisms that generate the patterns. We need to integrate that with ecological observations. And also we have one minute left. We also need to make to make inferences about the about the biodiversity. These are those analysis need to be performed at an appropriate scale. So I want to finish this presentation with this beautiful schematic figure from the paper from Jennifer Bender bodies that shows that how we can integrate multiple dimensions of biodiversity. But as well models and tools that allow us to to understand the biodiversity patterns and also to a conservation and management action facing global change. Thank you. Thank you very, thank you very much. Okay, that's great. So the next speaker sorry I got to is Christine Wilkinson from the University of California Berkeley Berkeley. Dr Wilkinson is the conservation biologist carnivore ecologist and science communicator at UC Berkeley and the California Academy of Sciences. Her research interests include multi disciplinary mapping human wildlife conflict carnivore movement ecology and using precip precipitatory methods for more effective and inclusive conservation outcomes. So thank you very much Christine will look forward to your presentation. Thank you. Can everybody see my screen and hear me. Yes. Awesome. So I'm going to be giving you a really broad introduction to participatory mapping which is one of the tools I use to prioritize equity in my research and conservation work. And let me just move this thing off the screen. I think you guys can see that toolbar. All right, here we go. So participatory mapping is essentially community members expressing their spatial knowledge about what is happening on the ground. So first recognizes that they are the experts about their space. It can also elevate and prioritize people's views and values about where they live so mapping is inherently political. It's usually done top down and it's often used to make very important decisions and this brings it back to being bottom up. And in so doing it could contribute to equitable decision making that recognizes various ways of knowing. So why should we use participatory mapping to foster more equitable conservation and human wildlife coexistence efforts. Well, we know that a wildlife co occur with people and even the most pristine quote unquote landscapes. And we also know that interactions between people and wildlife can shape ecology at many scales. And these interactions can also influence and impact people's livelihoods and health. And there can be so many drivers of human wildlife interactions. Many of them are multifaceted and can have deep historical roots driven by human culture and politics. And I've included this figure from a paper by Rahima White and colleagues that inspired me early in this work. Essentially we're recognizing that conservation conflicts are not solely the product of ecological elements on the landscape, but also social and economic factors and political histories. All of these are integrated together. And participatory mapping can basically help us to unpack all of this complexity and understand how human experiences interact with an influence ecology and animal behavior, as well as the other way around while also providing space to co create conservation solutions. So for those of you who are new to participatory mapping there are many tools out there both physical and digital from paper maps to digitizable paper maps to purely paper digital maps, and the best of these tools center equity. For example, this platform called Sapelli that is specifically designed for non literate participants to be able to report wildlife crime using symbols rather than words. And the ethical considerations for participatory mapping are arguably the most important part when you're thinking about it. So first two is involved what's represented in the room. What are the power dynamics that are going on and how can we mitigate and plan for those where and when are the sessions being held for accessibility. I like to think about bringing the table to the community and not just bringing them to the table. And how is the community involved given your funding and other sorts of logistical structures and challenges. When does the community get to buy in are they developing the project from the very beginning are they able to iterate upon it. How are you creating it together and being sure to be open and transparent about that from the beginning and throughout. And then most importantly, positionality and relationships are very important. So understanding who you are in relation to what's going on in those mapping sessions. And for instance, when I was doing this, I was working with two Kenyan master students. We were all students and we made that very clear from the beginning who we were. And we also would not have done this work. If we weren't able to and planning to maintain those relationships for a very long time. So onward to our briefcase study. I'm a wildlife ecologist by training and my research was helping me to understand the dynamics of carnivore ecology and movement in the crew County Kenya. But we also wanted to understand whether the ways we are characterizing these ecological patterns on the landscape. We're different from how community members perceive those patterns. So some background, our project is located in the rift Valley of Kenya in and around to adjacent protected areas. Lake Nukuru National Park and so I saw the conservancy and these two protected areas to some extent function as ecological islands. And to give you more context to this photo is a great example of the subdivided agricultural landscape surrounding the two protected areas. Also, so I saw the conservancy pictured here is not just a wildlife conservancy. It's a cattle ranch and it's been so since the colonial era. And this might look like a photo of the military, but it's actually the Kenya Wildlife Service, which is militarized largely to stop poachers. And because of this dynamic and the region's colonialist history, the people I know in the surrounding communities have a very complex relationship with the protected areas. And lastly, this electrified semi permeable perimeter fence around the National Park has become a complicated actor in the area with social and ecological winners and losers because of its shifted shifting permeability and stated purposes. And the main goals of the project overall, which I started in 2018 as part of my dissertation are to understand the social ecological drivers of carnivore landscape permeability, human carnivore conflict and coexistence and to use inclusive science and community engagement to shift the needle toward equitable human wildlife coexistence in this region. All of which is even more pressing with the increasing climate induced droughts in the area and focusing in on the parts where we used participatory mapping the most. We compiled data on verified livestock predation by carnivores and what I mean by verified is that, for instance, someone who experienced an attack on their livestock might call Kenya Wildlife Service and then KWS has a process to collect evidence and verify which animal was attacked and by which carnivore species. And of course we also use participatory mapping with 16 major villages outside of the protected areas. We had nearly 400 participants from these villages with about a dozen participants in each session so that the completed maps that they were drawing on would be legible to us later. We had to use paper maps because of our logistical challenges that we had mapping groups were also gender stratified to encourage people to speak openly with one another remember power dynamics. And importantly, we made sure that participants knew we were students we didn't have any particular sway over conflict management issues and the sessions were voluntary. Because a lot of people had not viewed maps before map orientation was really important. Community members were oriented to the map using an atlas by asking them to work together to point out places of importance or interest before getting started. And then they participated in one on one interviews and they were assigned a unique pen color for drawing on the map so that we could associate their interview data with what they feel is important in space. And you're asking them questions about their risk perceptions about carnivores livestock attacks and attitudes toward carnivores, as well as where they'd like those animals to live on the landscape. Lastly, we digitize the maps using field papers.org which uses the QR code and those reference points to georeference the photo of the completed map. So some of the stuff we're getting out of this is that the participatory mapping has helped us to see the disparities between people's experiences or perceptions of human carnivore conflict and what we're logging with the government records and with the ecological data. For example, on this map here the red areas are places where people have drawn on the map that they've experienced a lot of conflict, but the government records don't reflect that and the opposite holds true for the blue. Now that's already helped us to better target regions that we're not that we're not often reporting their conflict experiences to the government, largely through lack of access or understanding of conflict reporting avenues. And we tried to look a little bit at what local trends might be driving the difference in that perceived and verified conflict so we used geographically lead a logistic regression and saw that there were local regions where certain perceptions might be influencing an excess of perceived conflict reports. For instance, on this map I've circled two regions where the perceived conflict is predicted by these participants beliefs that their children face threats from carnivores while traversing those landscapes. Managers can actually use these local clusters and their predictors to decide on targeted outreach for these particular communities. For instance, perhaps the fence is being breached more in these areas and it's not being picked up by our ecological techniques or more transport can be provided for children to get to school and avoid these real or perceived risks. And briefly we're also integrating the participatory map data with GPS data from a major conflict prone species in this region the spotted hyena. And that's allowed us to build models that show how social ecological factors predict hyena movements and subsequent interactions with people. For example, at a landscape scale ecological factors are just as important as human elements, but at a finer scale human factors are more predictive than ecological components. So some case study takeaways. First there are quantifiable spatial mismatches and verified and perceived conflict and there are ways to predict these risk perceptions and the attitudes that influence them, meaning those are great jumping off points for addressing conservation conflicts. Also understanding spatially explicit context dependent risk perceptions can improve targeted community engagement around conservation. And most importantly, integrating these disparate data streams and using inclusive spatial planning with diverse and representative members of communities can help with disentangling conflict, elevating community voices and ideally co creating targeted coexistence measures. So main takeaways about participatory mapping. First mapping is inherently political community participatory mapping can get you involved in tough conflicts around local boundaries for instance, if you're not hyper aware of who is involved in those sessions. Second back to positionality really really important. For example, while we told folks we were students at the time. We also made them aware of the power we did have like being a bridge between the community and the government agencies that make a lot of decisions in the area and that was really helpful for community members. Relationships are very important. We would not have connected this work if we didn't intend to be embedded in this region for the long term. And lastly, we can talk about this in the discussion scaling participatory mapping efforts and results can create really difficult challenges, but there's a hopeful note, and that's that the community engagement and empowerment that can come through participatory mapping can be just as important for conservation as the data themselves. So I'll just say thanks to all these folks funders etc, and thanks y'all. Thanks Christine that was really great very thought provoking thinking about that interaction between people and animals and conservation and how they anyway, we'll get that in the discussion. So our last speaker today is Dr. Elise Zipkin, who is an associate professor at Michigan State University and she is a quantitative ecologist connecting the complexities of natural communities with the precision of mathematics to shine lights on mysteries of ecology and conservation. Dr. Zipkin and her team developed analytical frameworks to address grand challenges in the study of biodiversity loss and the effects of anthropogenic activities such as climate change. So thank you very much Elise and we look forward to your presentation. Sorry I realized I was muted and I'm having a hard time unmuting myself. Can you see the screen. Yes, we can. Thank you. Thank you. So, oops, I don't know what happened there. Let's start at the beginning. Okay, so what I came to talk to you about today is data integration and I feel very lucky to go at the end of these two sessions on the research tools bottom up and top down approaches because I think people hit a lot of kind of what I want to talk about today. And really I'm going to go into the challenges related to data integration but I want to start off kind of getting us all on the same page here with that. So first, the idea of data integration kind of comes in a lot of different names and people have been talking about that today in very many different ways but I'm really talking about the idea of integrating multiple data types into a single analytical framework. And that's different from things like data harmonization, where we're taking this different data and trying to structure them in some way to put them in the same database, or it's also different from other approaches where we use different kinds of data with to answer a question but not necessarily within the same particular model so that's really what I'm going to be referring to and I talked today about the different things that that the challenges kind of associated with that. So why would do we want to do integrate data integration. Again, I feel lucky because so many people went before me to really highlight the importance of combining different kinds of data types but the big kind of reason for this is really that we can get increased accuracy and precision in any kinds of biological parameter estimates that interest us. So this is just some an example here where we were integrating two kinds of data, a presence only data maybe that's only where we see opportunistic sightings, and where we use some sort of more structured data with a sampling design, and we can find that we can get better estimates. What's the effects of different covariates like the landscape on the presence of individual species, and also we're able to estimate kind of the overall abundance much better by integrating those. So there's a lot of different reasons why we might be doing data integration, including also being able to estimate parameters that we can't with individual data sets and look at all sorts of forecasting and other kinds of effects in the future, but this is sort of always kind of the main one. So why do we want to do data integration for macro systems so macro systems and I'm sure this is just repetitive at this point is the study of ecological processes and patterns at really broad spatial scales, as well as their interactions with phenomena at other scales so the kinds of research that we're talking about at a continental scale. So the reason for this right is because we're now thinking about problems and processes across multiple scales. And what we often find is that different kinds of data are collected at different spatial scales. So if we're thinking about trying to collect data across a really broad spatial scale something like you know the continent. So we end up finding not always, you know, there are obviously different kinds of examples for us but a lot of times we get, you know, either sparse data so we're only able to have information at some particular sites for example maybe neon, like where we have a lot of content but we don't necessarily have everything in between filled in and or we get low information content data so what I mean by that is kind of unstructured data maybe we're lucky we have some sort of public science programs where volunteers are going out recording organisms that they see that kind of thing but there might not be any design with those data. So really high information content data the kinds of data that a lot of people talked about here generally although not always tend to occur at smaller spatial scales it's a lot more effort to go out for example, and capture and recapture individual, you know individuals of a species to record what's going on with their reproduction their survival all sorts of things so generally we find those to be at larger at smaller spacer scopes spatial scopes but those kinds of data have a lot of information in them. So the benefit of studying kind of continental scale questions is that we can combine these different kinds of data types, and they can give us different pieces of information about what's going on spatially or about processes into a single analytical framework. And now I'm getting into the meat of what I want to talk about, which is the challenges of data integration. So unfortunately I don't have time to go all through this but a group of us wrote a paper who work on macro systems research that was published in New Jersey College environment a couple of years ago in the special issue coming out from the National Science Foundation on macro systems biology research, and it just kind of goes through a lot of these particular issues this is just an, this figure on the left is a graph showing how much more data integration we're seeing the number of studies that we're seeing through time so really this is something that's increasing. The big one is resolving spatio temporal mismatches of available data. So other people talked about this too but you know there's kind of the idea that with remotely sense data versus something like collecting data in situ, you tend to collect that information at really different spatial scales. And that can be a challenge when you want to analyze those all together. So here is an example where different data were collected on black loaded blue warbler across Pennsylvania at different spatial scales. So one, one data type was collected at like a broad spatial location, you know, five by five kilometers and others were collected at a smaller location. And so kind of if you naively put these together, you can get you know you get this map of what they got if we just kind of don't think about this spatial problem. However, when these this group of authors work together, and they incorporated the different processes that were happening at the spatial scales through what's called change of support processes, you can get a much finer resolution of what's going on and a more nuanced understanding of okay the birds aren't just all in one place, where within that do we see what's going on. So, you know, if we appropriately apply these kinds of methods, we can get better understanding of what's going on, oftentimes at a finer spatial scale. Another issue is the idea of unbalanced data so that just means, you know, one data source has a lot of data points, while another may have very few so the example here is, if we've got some kind of automated system on the like, like what we see on the left there, we can be collecting data, maybe at the order of seconds or less than seconds. And, you know, the contrast that with what something where people are going out these are both actually at neon sites, where people are going out and collecting data. Naturally, there's there's going to be a much smaller number of data points so different very, very different volumes in those data. So this how to deal with that is a really active area of research right we could always kind of thin our data of the more volume heavy data. But again, that gives out, you know, may take out some information that's important. And also we can, you know, what what kind of people are now thinking about is formally modeling those biases that cause like the auto correlation, or the different information in there into the model itself, which can account for that without having us to maybe throw out some of the data that are in there. Another one that we find a lot in public science or volunteer based science is accounting for spatial biases and data sets. So this is an example on the left here this is a map of Ohio. And these are locations that people have gone out to collect data and what we find here as we often find is that people go, you know, collect data or near their homes when you're asking volunteers where to collect to collect data, they're doing it near where they live right they're not creating a stratified random sample or some sort of other random sampling approach that we would particularly generally want to use as scientists in the field. So there's this can create a problem right because if they're going or the people often also go to areas where they think they'll see stuff so this is you know for looking at butterflies well nobody, no volunteer wants to go out and collect data where they're not going to see any butterflies right but that information is really important to scientists and we often do go to areas. You know where we're not expecting to see so many because we want to know that as well it's really key to understanding you know what's going on. So there's a wide again wide array of solutions for how people have been thinking about addressing this problem. And you know, a lot of that relates to developing hierarchical structure to models, incorporating spatial random effects that sort of thing, and kind of other approaches similar to the one that we just talked about is to incorporate this model component directly within the model so we can add the site selection process within our modeling framework that's sort of newer approaches to thinking about that. And then another issue I think that that we're starting to think about more but is kind of you know not not been around quite as long as the idea of non stationary and basically the idea of that is the effect of some variable let's say weather or climate in one region may not be the same everywhere else it might really be different. And so because of that we can't assume that everything's the same so it's a variable that stationary is the same everywhere. And then other kinds of non stationary variables might mean okay the mean is different in one place or neither weather effects, let's say rain effects, you know how much we see butterflies differently in one region versus another because of some other unknown factors. One minute left. Okay, and then the other might be that we also see changes in variance for this. So there's lots of different ways that we can accommodate this we can model what's going on differently for different areas, we can incorporate hierarchical random effects, we can index these parameters by space and time that allowed it to vary. So I just want to wrap up by saying, you know that there's really quite a lot of opportunities in data integration because, you know there's thousands now of citizens projects, and it's the volume of data is just so so important for us for thinking about large scale continental kinds of research. And so I'm excited by the opportunities that we have out there and hope that this may be inspired some people to think about working on some of these challenges so we can answer these really important continental scale questions. So thank you. If I'm going to figure out how to stop sharing my screen. Great. Thank you very much. So now we're open to the question and answer period and there's a question from Slido for Nico. Great about the five years of funding for symbiota but what is the plan for long term data curation to ensure quality and retain user access to these data. Thank you. So it is a sustainability grant, which means that in the spirit of the National Science Foundation. We actually did a business design exercise. This is one component of our sustainability, but I'm just pasting it in here. We actually created a recharge center at ASU. We can charge fee based services for about six products that we have identified. If you can click on that. I'm just going to read it off. So we can charge. So this requires having good boundaries between what is in scope for a particular federally funded project and what is out of scope for that project. So understanding different clients to what extent they're engaging you in a very research heavy versus more of a service heavy environment. And also bringing the community along with the notion that, you know, good services to some extent require a community that self values. And one more aspect, bringing folks along with a, an idea of pay as you can. So there's a number of factors involved. We do have a bit of a in road because part of the symbiota sustainability is tied to the neon project, which has a projected value for another 26 years. And based off that, we are able sort of like to stack things on top of that, hopefully, but but it's, I would say it's, it's what most of my research nowadays is focused on is to actually build that sustainability. Great. Thank you very much, Ines. Thank you for all the presentations on all your insight. So as we have been listening to some to all of you. It seems to be that we have three major challenges. What is about data that we need more data, especially more in situ data. The other one is, maybe we haven't heard about this much this afternoon, but definitely this morning about data processing and cleaning the data making a useful. And now we just heard from the ladies about integration models. So, which one of these three, even all of an important but is one of them definitely bottleneck to allow us to progress with these continental scale ecology. Thank you. I think actually, one of the big issues that maybe affects all three of these is kind of computational power and computational time. And maybe because of that, that's kind of something that we can be thinking about that would probably advance, you know, at least two and you know, to some extent, but even potentially even one if you think about that with some of the data that are collected autonomously right so I think something that I see run into particularly with two and three is there's this issue of, you know, not being able occasionally to be able to, you know, quickly clean and process and handle the data and then also be able to model it together so I don't know if that exactly answers the question. But it's something that came to mind that we could probably work on together as team and move forward maybe multiple facets of that. Anyone else like to chime in. Just. Yeah, thank you just to put something out there. And of course this is only one person's opinion. I think a good part of the, and I am not by training a conservation biologist, more of a biodiversity data scientist, but a good part of the academic culture that I have experienced and have grown up in is not service oriented to necessarily be able to respond to, you know, societal demands in the ways that are as as effective and as nimble as I think given the way that our global environments are developing would be necessary so I think focusing more on services and focusing more outwardly is is an important bottleneck. And Jesus you had a comment. Yeah, I agree with it with the lease on Nico. But I think that the. All of the end is everything's in the will ended up being a bottleneck. Every prediction will depend of the quality of the data. So you can have the perfect model. But if your data is not, it's not good, your model will not work on the reverse as well. So you can have like a perfect data but you're not your models creepy so it will work. Thank you. I have a question. And so this can be targeted to you at least sort of how partnerships between community and researchers and then local stakeholders can be fostered to support implementation and sustainability of continental scale continental scale initiatives. Yeah, that is a great question and I think I'll also turn it over to Christine when I finish. Thank you. The only thing I would say is, you know, I think about this a lot now because a lot of the data that we use, you know, is from volunteer scientists. And I think sometimes there's a mismatch right there. They don't necessarily know what happens to the data afterwards, or how important is people like, I've noticed a lot of people do it because they like doing it which I think is a really good reason. But you know, we're trying to work more to come back to those communities and, and, you know, really provide provide results like and tell them what we're doing and even if we don't have all the results you know showing them the updates of the kinds of things that we're working on hearing from them like what are the challenges what they like doing about that, even having researchers go out with them sometimes you know that's something that I've been able to do and it's always really fun and I think it it shows people. So today, you're really an important part of this, you know, process. And Christine I think your perspective would be really useful on this question. Yeah, I'll just add to what Elise said by noting that a lot of the volunteer scientists come from certain demographics, a lot of the time, and we're missing out on like large swaths of not just demographics but also locations. And demographics tend to live where we have very little idea of what's going on, as far as biodiversity goes and so. Yeah, I like the idea of making volunteer science a little bit more participatory and intentional, because we have these amazing tools they've been rolled around the landscape and like I naturalist etc. And they've just been allowed to proliferate on their own but kind of going back and looping in what we know about who's being left behind. And then the other thing I wanted to add to that is we need to incentivize that type of engagement like the academic structure does not incentivize that engagement it pretends to incentivize that engagement but it evaluates you on one thing which is publications. And so once you get embedded enough within the ivory tower, and you become overwhelmed with everything that you're required to do and the one thing you're evaluated on. It's going to be very difficult to stay nimble and to stay community involved at the same time so I think changing the incentive structures is really important. Yeah, it's a great point. Bella. Yeah, I have a question is something about about something you said Christine about positionality and and acknowledging positionality in the research and I was just thinking about how positionality. It affects or creates these data gaps, and it affects how we integrate data it affects how we interpret data but these positionality statements are, are not, you know as an ecologist it's not something I've ever done or written in my traditional ecological papers. I think it can be really important, you know, especially with some of these data products so how could we, you know, could we do that, you know, include positionality in this type of work or and how could we incentivize that. This is a fantastic question and I'm only going to touch the surface and only from my own opinion and so one thing that you just touched on is that we as people train, I think many of us have been trained in the western scientific tradition, have been taught that objectivity exists when it does not right so all of your funding and the way any questions you choose to answer and the way that you choose to answer those questions come from various places that are subjective. And that's why we've been talking a lot about recognizing different ways of knowing today, right. And so I think that there's been a lot of exciting work on the AI and sort of data science front, recognizing how algorithms have been replicated and produced that have certain positionalities behind them that make them harmful over time, and trying to figure out how to mitigate that going forward especially given the use the increasing use of AI so I think that's like a really nice potentially extremely impactful movement that's going on even some folks at Berkeley I think are working on that. I think there are other ways to do it on more of a grassroots level to recognize positionality so like specific institutions embedded in specific places can think about creating long term community. Sort of collaboration efforts that that work with community organ community based organizations for the long term to both develop and answer research questions and co create that process together. I think that happens in some places but not everywhere. So those are just two thoughts that I have about that. Thank you, Stephanie. Yeah, I'm thinking about workforce development needs, and I thought this would be an interesting group to ask because you've got all of you have talked about data integration challenges but you're also representing some of the collections challenges and, you know, the needs for tools and knowledge to work with communities closely. So, I, that's a very open question for anybody who wants to jump on it. I guess I'll go with that since I'm also the director of our ecology evolution and behavior program so I think a lot about training. I mean, one, I'm biased. So the bias that I have is, you know, that the quantitative methods will always be serving us well regardless you know if we think about. I'm going back to that bottleneck question you know the collection of data the processing of the data the analyzing the data, really that kind of analytical thinking I think quite a bit helps design each of those. I mean one of the things that we're talking about you know with the volunteer science is the idea that we're going to get data, we're not going to have control of maybe the data that we're always getting. And so the analytical skills I think can be really helpful for thinking about all right what kind what is the useful data that we need to actually get to answer the question that we have. But then I think even more broadly than that is really important the communication so you know that's something we're thinking a lot with our students. Communicating why this is important to everybody that's involved to every stakeholder that might be, you know, interested in that I think, because, you know in the past it would just be like scientists working by themselves and producing whatever but we are facing really big important challenges, and it's important for us to be training, you know the next group of individuals that's coming to be able to talk about why what we're doing is really important when they are million competing things going out in the world so I kind of think those two sides are two of the things many things that people need to be successful in this but that's my two cents. We have time. I'm sorry. Do we have time for others to. It'd be great if anybody else could chime in on that question especially maybe a Christina Jesus since you're coming up through the ranks. I feel like I'm like the depressing one though, so I do feel like there are quite a lot of since I'm talking about community science and participatory work and objectivity subjectivity ways of knowing there are quite a lot of young scholars that come to me to ask for advice about how to incorporate that kind of work in what they're doing and how to stay community engaged despite all of the challenges in doing so and by challenges I mean like funding structures etc that don't support that. Like my advice of like little back doors that they can take and ways that they can approach it ethically and all of those things which are important to think about. But in the end, like often you run up against the systems themselves that that sort of stop you from taking all of your ideas and well meaning ethics and so on to where they can be so. I do struggle with answering that question from the community engagement side of things because there's a lot of eager folks and a lot of great trainings but then a lot of folks kind of fall off because of the systems. Get the best of them and suck their soul out of it. So, I don't know what you think Jesus. I think that those are great, great point Christina and I will I will read you for advice. But yeah, in my case is that I didn't get my PhD in the United States and I did my PhD in Brazil. I think that navigating this different cultures and perspectives is is is is is difficult but but then when when you get there, you can continue forward, but it takes time. I think that the as Christine pointed is when it grows that support that indicate that in other words like taking it from your hand, take your hand and guide you in the process to get there. I know that makes sense. That's great thank you and Christine I think that's kind of a good news bad news, you know bad news maybe the system hasn't changed but good news that people are coming to you to ask about that aspect and that's really great news actually, because that means there was kind of a boiling of of young people that are interested in using those technologies and techniques and realize that they're bigger than themselves in terms of what they want to do. Thank you all for your time.