 Hello, I'm Charli Well, a data scientist at the Natural Capital Project at Stanford University, and I'm here today to introduce the Invest Carbon Storage and Sequestration Model. This is the last model we'll learn in this training, and you'll see that it's much simpler than some of the other ones we've seen this week. However, it's still a very useful and widely used model. Indeed, terrestrial ecosystems store more carbon than the atmosphere, so they are vital to influencing climate change. Terrestrial carbon sequestration and storage is perhaps the most widely recognized of all ecosystem services. The social cost of a ton of CO2 released into the atmosphere is estimated somewhere between 10 to 100 U.S. dollars. So this model is used in many decision contexts. For example, in the context of red credits, red or R-E-D-D stands for reduced carbon emissions from deforestation and degradation. So in the context of red credits or other payments for ecosystem services schemes, knowing which part of a landscape store the most carbon would help governments identify opportunities. So opportunities to earn red credits or efficiently target incentives to landowners in exchange for forest conservation, etc. Other examples of use may include, say, a conservation NGO that wants to invest in areas where high levels of biodiversity and carbon sequestration overlap. So the outputs of this model showing where carbon is stored on the landscape would also be helpful for them. Or similarly, a timber company may also want to maximize its return from both timber production, but also red carbon credits. So this model is used in several decision contexts. But how does it work? The invest carbon model is very simple. It follows the IPCC inventory approach to assigning carbon storage values by land cover classes. So in four different carbon pools, which we'll see in a minute. So the model can estimate carbon storage, which is the mass of carbon stored in a landscape at a particular point in time. But it can also calculate carbon sequestration, which is the change in carbon storage in a landscape over time. And optionally, if you want to assign a monetary value to carbon sequestration, the model can perform valuation. So the model estimates carbon stored and sequestered in four carbon pools. The first one is above ground biomass. It comprises all living plant material above the soil. So bark, trunks, branches, leaves. The second one is the below ground biomass, which is the living root systems of the above ground biomass. So all the roots. And then soil organic matter, which is the organic component of soil. And this represents the largest terrestrial carbon pool. And finally, the dead organic matter, which includes litter as well as lying and standing deadwood. So carbon storage is simply the sum of these four carbon pools of the carbon stored in these four carbon pools. And then carbon sequestration is the difference between storage at a certain point and storage at a previous point in time. So what are the data needs of this model? So it requires mainly two inputs. The first one is a land use land cover map that we will shorten by LULC. And the second one is what we call a biophysical table, which is a table with carbon stock values for each land use land cover class. And optionally, if you want to calculate sequestration, then you need a land use land cover class for another point in time. Or if you want to perform red scenario analysis, then you need a red policy map. So another land use land cover class with a scenario of implemented red policy, as well as economic data, if you want to perform valuation. About economic valuation, as I said previously, this is only done on sequestration. It's also based on the net present value. And to perform economic valuation, you need a couple additional model inputs. So you need an estimate of the social cost of carbon, which is the value of a sequestered ton of carbon. Depends on the economic context. You also need to estimate the market discount rate, which basically captures what's the preference for immediate benefits over future benefits. Also depends on the economic context. And you need the annual rate of change in the price of carbon, which adjusts to basically how much carbon emissions will impact climate change in the future versus today. So it reflects changes in the societal value of carbon sequestered. So now the outputs from this model. The main and most important one is the map of carbon stored in the landscape. You get one carbon storage map for each point in time for which you provided a land use land cover map. If you gave several points in time, you'll also get the carbon sequestered, the difference, as well as sequestration maps for red scenarios analysis if you provided red policy maps. And if you provided all the inputs for economic valuation, you also get a map of economic value of the carbon sequestered in the currency of your choice. So the currency in which you gave the inputs. And the model also provides an HTML summary of results. So this is a very simple model, which has a few important limitations. First, it assumes an overly simplified carbon cycle. Also, it uses the IPCC static inventory approach, which makes a lot of assumption, including the fact that it's assuming every hectare of a given land use land cover class is identical. The model also assumes linear change in carbon sequestration over time, which is known to be wrong. That's the figure on the right here. And potentially it assumes inaccurate discounting rates for the economic valuation. Some of the biophysical conditions that are really important for carbon sequestration, such as photosynthesis rates or the presence of active soil organisms, are also not included in the model. So overall, this model is greatly used in a great, fast way to estimate carbon stored and sequestered in a landscape. It is widely used because it's the IPCC method and it's not very hungry in terms of data needs. It's pretty easy to run and you don't need a lot of data. But it's important to keep in mind that, although this is one of the best available methods out there and so easy to use, it is not exhaustive and it is not a real reflection of reality. A couple important resources, the user guide. As for all invest models, the user guide is really your best friend. It contains all key information about the model and all the underlying equations as well as references to all the scientific publications on which the model is based. But it also contains useful data sources, both local data sources and a lot of global data sources that you can use to parameterize in particular the biophysical table. And in the forum pointing out again, community.naturalcapitalproject.org, the forum is really the go to if you have any questions or bug on the model. Please look if someone had your question in the past because often that happens. If you have a weird error message or something bizarre about the model, definitely try to type it up. This model has been used for several years, so hopefully it happened to someone before and the solution is already out there. All right. So now let's give it a try and see what it looks like running the model. So I am opening invest and it's the first model on the list. So launching the carbon storage and sequestration model. So then you have to choose first a run space, so a workspace, sorry, so where the model outputs will be run. So let's create one test run space, for example. And then you have to choose your land use, land cover map. So here I'm just picking mine, which let me show you actually in QJS what it looks like. So we are here in the eastern step of Mongolia. Let's zoom in a little bit. And my land cover map looks like this. So this is a land cover map that we also provided for you. And this is, it's derived from MODIS. Yeah, it's derived from MODIS. So this is my land cover map. The next thing I'm going to provide is the biophysical table. So my biophysical table looks like this. It has, for each of the land use code that are encountered in the land cover map, it has a value for each of the carbon pools. So in that case, this is what I provided to you for the exercise. It's a very simplified version. And I just focus on one carbon pool, the above ground carbon. So I just set all the others to zero, so I don't worry about that for now. So once these two inputs, which are the only mandatory inputs are set, then I can just press run and run the model. Once the model completed successfully, you can look at the outputs. So you will have total carbon current, since I just gave one point in time, which is current. And then you have the results detailed for each carbon pool. Because I just provided data for carbon above, that's going to be the only meaningful one. So let's take a look at it. I hit style it here. So this is in megagrams per hectare, megagrams of carbon per hectare, the output of my model. And I can see, for example, here that there's a lot of carbon stored in this area, which is easily explained by the fact that this is the most forested area on the landscape. One thing that you can do once you've run it, if you have external data sources of maps of carbon stored above ground, which I have this one, but it contains a lot of no data in the desert, is you can compare your datum to other external sources so you can validate the results. So that's about it. That's all I have to show you today. Good luck, and we look forward to see what you make with this model. Talk to you soon.