0:13 Verena Vogler, a new ecological simulation framework for Rhino Grasshopper. But first of all, yeah, I will say a few words about our two main products, which are Rhino and Grasshopper. So, maybe question to the audience, how many of you know Rhino and Grasshopper? Just some hands up. Okay, good good to see. And how many of you have opened Rhino and Grasshopper in the last week? Not over Easter, but okay, good good to hear.

0:43 Yeah, well, so Rhino is a free-form modeling tool allowing precise control over complex shapes and forms. And Grasshopper extends Rhino’s functionalities by providing a parametric design environment and algorithm editor that is integrated, and it helps you to create your own computational workflows. And both tools, they are developed by McNeel. McNeel, a few words about the company, is a privately-held and free-owned company that was founded in 1980.

1:16 Our headquarters is located in the US, in Seattle, and we have regional offices worldwide. In fact, our European office is approximately 10 minutes away from this venue, where Yeah, based in Barcelona, near Arc de Triomf. So, if anyone is here wants to say hi at the office, you’re welcome to pass yeah, to come by. And maybe a thing that’s important to say, it’s compared to other software vendors, we are really actually we’re a relatively small company with only 130 employees worldwide.

1:52 What we do is we develop our own public SDKs, and also in-house geometry kernel and also 3DM file format is open. Yeah, it’s freely available and open since more than 20 years. At McNeel Europe, we have an R&D team and since now it’s over 10 years, we have participated in Horizon 2020 research projects that are funded by the European Commission. And in these projects, we try to explore our new frameworks, open APIs, and new fields.

2:26 So, we have worked in the past with Yeah, in fields like geomatics, neuroscience, and lately we started to work also in a field that is not so familiar for us, which is ecology. And then from this work, actually this new tool Rhino Ecological emerged. So, let’s start with the problem that we had and why we thought that would be interesting yeah, way to approach bringing this simulation to Rhino.

2:57 So, the problem was in most design workflows today, ecology is treated as something static, something that we add at the end. We assign planning schemes, define these green areas, or run environmental checks to quantify parameters such as biodiversity after design is already defined. And actually, we see this in the industry, there are a lot of new standards where people need to prove when they Yeah, submit a project that they kind of need to also submit these indexes, biodiversity index, and so on.

3:32 Yes. So, this tool is also addressing this. But living systems, well, they don’t work like this. They grow, compete, adapt, and change over time. And that dynamic behavior is almost completely missing from how we design. Furthermore, the design disciplines, and here we talk about architecture, urban design, landscape architecture as the tool is a design for those. They considered spatial because their primary medium is organization of form, geometry, and relationships in space.

4:09 So, ecology, however, is spatial and temporal. It unfolds over time through growth, competition, and also succession. And this gap between space and time is the core challenge when it comes to design with ecology. So, Rhino Ecologic addresses this by combining architecture, ecology, and computational design. It tries to link these kind of these three domains, which is geometry defines environmental conditions, environmental conditions define that is ecological processes.

4:41 So, instead of assessing vegetation, we simulate how it develops. It outputs the pool outputs location and time-specific 3D distribution maps as well as biomass and biodiversity analysis. So, here you can see at the computational framework of our system. The So, in the first step, design geometry is converted into a voxel-based spatial domain. And to this domain, environmentally environmental conditions they are mapped. And these include solar exposure, soil conditions, and also precipitation.

5:26 And then these conditions, they drive ecological simulation where plant communities can evolve over time. And finally, the system it outputs a lot of data that can then be evaluated and fed back into design process. So, the process could also be like a loop. And what is at the core of the system? There is a spatial discretization, which is yeah, the geometry, your Rhino model, your project is discretized into a voxel grid, where each voxel represents a localized portion of space with its own environmental properties.

6:08 So, to each voxel a center point, which is XYZ, we map property values V, which are attributes for geometry, environmental aspects, and also ecological aspects. And they’re then connected through a graph data structure, so each voxel cell is a kind of aware of what’s happening in the voxel cell and a neighboring voxel cell. I would like to introduce you now to the environmental models that are running behind Granny Ecologic.

6:43 So, environmental conditions there derive directly from the geometry and location, and this includes solar exposure, soil depth and volume, and precipitation, and it’s computed spatially across the voxel grid, rather than yeah, a simplified averages over the whole model project. So, solar exposure is calculated based on the geolocation using a set of equations and mapped onto the voxel grid, allowing us to evaluate both daily and hourly sunlight.

7:16 In the soil model, it simulates how soil accumulates across the site. By default, we use this kind of snow form model, so everywhere where snow can fall, where it could have potentially soil, that’s the default setting, but the user can also define where you would like to have soil, for instance, on vertical areas, and so on. And also we have a reset of soil types, but the user of course can also define the soil types that you have on the on the site.

7:49 And the precipitation model, it calculates how much rainfall the model receives based on and this is based on open data that we access. So this means environmental heterogeneity is no longer abstract, it becomes like an integral part of the model. And before I would like to present you now you have an idea a little bit about the environmental models and before I would like to show you what we did and at the back end with the ecological model, a few words about what ecological modeling is.

8:42 So First there are more there are different models in ecology, computational ecology and then our team we have a computational ecology ecologists that is aware of these models. So first there there are these models like species distribution models. They are called SDMs and they predict where species can grow by evaluating environmental conditions such as temperature, precipitation, light and soil. But however these models they are typically limited to two different predictions of of expected plant presence and they mainly rely on abiotic factors and only partially include temporal dynamics and they operate at the course, they are really data heavy and they also operate on large scale resolution.

9:29 Another model that we look at we are looking at and it’s a typical model in ecological modeling, it’s called habitat suitability. And that describes in a way how much suitable how suitable a specific location is for species to grow based on environmental conditions. And in practice, this often gets simplified. This kind of this habitat suitability model heavily relies on soil. My factors like light are highly variable, and precipitation is also not consistently integrated.

10:08 So, it tends to explain how well something might grow rather than why it might fail. So, for designers, that means soil becomes the most critical layer because it directly controls water and nutrient availability, setting the basic conditions that determine whether a species can actually establish and survive in a certain space. So, our colleague is called Uschinski. So, we integrated the Uschinski model on Rain Ecology, and that works slightly different.

10:41 So, instead of describing where plants could grow, we simulate how they actually live. So, with the Uschinski model, every plant is treated as an individual agent, and these agents are grouped by species. So, for example, all daisies and this plant species follow the same behavioral rules, but each plant still exists and interacts with its local environment. And what makes it powerful is that plants don’t just respond to conditions, they interact also with each other.

11:13 They compete for light, they grow according to species-specific biological rules, and follow real-life history dynamics like lifespan, height, and reproduction phases. So, growth is then not as linear. It follows curves like the Gompertz function, which is based on empirical data that you can find in scientific literature. And importantly, this creates this kind of emergent behavior. Things like shading, they’re not predefined any longer. They arise naturally from the interaction of many individual plants.

11:56 That’s something you can’t really predict. Also, if you use AI, you cannot really predict this. Mostly, you cannot predict this from static models or precomputed data. The model also integrates habitat suitability, light resources, distribution, plant shape, and seed behavior. And it brings all these factors together in this one system. So, we in a way combine habitat suitability with agent-based modeling. Yeah, linking environmental conditions with this dynamic plant behavior.

12:33 And from here, we can visualize how plants grow over time, how species distribute across a site, how biomass develops per species, and how biodiversity emerges in a specific design. So, instead of designing with assumptions, we can design now with the living systems. And this graph is one of the first tests that we did. We didn’t have a plugin yet. This was just a Grasshopper definition where we integrated all these tools.

13:01 And it’s going running really slowly, but this one of the really early test. Yeah. So, from here, how does this application look like in Grasshopper? So, a little bit about how we designed the concept. So, we start with a kind of project setup. We define the site geometry, and you voxelize it into the spatial grids. You can define the resolution and from there we run certain environmental models, analysis models and we compute factors I just mentioned and all this information is then stored in a data frame where each voxel cell contains its environmental properties.

13:45 So and on top of that we run this ecological simulation and finally extract the data. Such as species distribution, biomass and biodiversity metrics. So the workflow moves from in a way from geometry to spatial data and to in the end also measurable outputs. So this is the how we have these functions integrated in the in the Grasshopper. So it’s just a layout about different functions that we have.

14:21 This is how the UI looks at the moment. I’ve said the plugin at the moment is in alpha. So we haven’t really publicly announced this. But yeah, so on the left side we have a whole yeah, we have new parameter types then we have components to set up the projects. We have components to analyze the projects. We have components that manage your data. We have statistics components and we also have components to visualize your data and if you want to create a custom species library for your site, you can also create a set up your custom species library.

15:01 And just here a small example project with a simple yeah, geometry. Here you can just drop a component, create your project, define a geolocation and reference your Rhino model but it can also be Grasshopper geometry. You reference it in yeah, in Rhino Ecologic. And in the next you just drop in Rhino Ecology analysis component and then we correlated all these different processes so you can kind of get results for environmental analysis, which is yeah, soil depth, soil volume, soil type, precipitation, solar aspect and as well the biodiversity parameters.

15:48 Then you can just drop a few Grasshopper components to visualize this data. So in this case, you have the solar exposure analysis. You get your soil information that you can visualize precipitation. And here this is the most interesting part. You can flip through the different time steps and see how the vegetation that you have defined in your model evolves over time on your project site. And there is also a growth simulation where you can see how it what actually how this actually works in space over yeah, couple of yeah, in this case of 30 years.

16:32 But behind these visualizations, there’s a lot of data for the people that are familiar with Grasshopper. You could just visualize data in Grasshopper by adding a Grasshopper panel. So here you can see that this is just for one time step that we have data for each voxel. So we know how what kind of species do live in each voxel. We know biomass values for each voxel cell for each species, the abundance, which you can compare to biodiversity and also we have in the end period, which is called plant volume for each voxel.

17:11 So maybe there’s a lot of information and there’s a lot of data. So what we also introduced in Rhino Ecology. It’s Yeah, it’s kind of a data analysis components that analyzes this amount of data because we have a lot of attributes in each voxel cell. So we have a machine learning model which is a Python model and this analyzes your data. So you can actually yeah, you can get these graphs where it shows for each species how the biomass evolves or the biodiversity of the site.

17:49 Or you can also get these kind of correlational maps, which is also interesting because then you can understand what’s actually going on in the data. So for instance, if you increase the elimination on the sites, you get 4% more in biomass. Or if you change the soil type on the terraces from soil type five to soil type nine, it increases as well the biomass. So you can actually use this data to adjust your design.

18:20 So we have also some users since the first prototype of Rhino Ecology, we have collaborated with the Architectural Association London. We’re now Rhino Ecology is part of the N Tech master program and the PhD programs. In this example, the alpha version was used to integrate environmental performance into the design of these agricultural settlements. And then based on these inputs such as solar exposure, site topography and species growth behavior, the system generated these density index.

18:54 And then this index was used for multi-objective optimization to compare this kind of spatial configurations and identify this design outcome that actually maximizes the ecological productivity. And this is just another example where something similar was done. And we also work together with Henning Larsen Architects for Copenhagen’s first all timber neighborhood, which called Falbe and also ran some studies for them. We have a few testers now. I think we have now over 60 different companies organizations that are testing right ecological.

19:40 Most of them are from the industry and also from A’s from A’s. So this is the majority of people that use it, of course. And then yeah, if you are interested also to be our tester, you can just yeah, take a screenshot and sign up as a better tester and then you can get our latest build. Well, a few conclusions. Yeah, it’s a little bit it’s quite important to clarify what the system is and also what the system is not.

20:17 It does not predict exact outcomes and it does not replace ecological expertise. And it’s a design framework a way to explore ecological dynamics within the design process. And this enables design approach in which ecological performance is analyzed, compared and directly integrated into design decision-making. And in this framework ecology is in a way not just an external constraint but an active system within the design space. So, that’s it.

Thank you. And if you want to learn more and and also if you have some questions concerning Rhino and Grasshopper, I’m here today with my colleague Carlos. He’s here sitting in the first row. So, yeah, just come over and talk to us. So, thanks a lot. To learn more about the CDFAM Computational Design Symposium, access the archive of previous presentations, interviews with speakers, and information about future events around the world, visit CDFAM.com.