Why is modelling land-cover changes over time so important and challenging?

It is now widely known that climate is expected to change over the coming decades causing dramatic impacts on biodiversity and modifying the suitability of the landscape for all species (Sage, 2020). But, climate is not the only factor affecting the distribution of biodiversity. In fact, one thing is even more important: the distribution of land-use and land-cover. They encompass the precise location of natural habitats such as forests, grasslands or wetlands, as well as managed areas such as pastures, crops and even urban areas.

Having access to this information is key for several reasons. First of all, it helps understand the vulnerability of species because land-cover changes are the main cause of biodiversity loss worldwide (IPBES, 2019). Second, each class of land-cover has specific characteristics in terms of ecosystem functions and services, also called Nature’s contribution to people. For example, forests or wetlands have the capacity to store atmospheric carbon which participates in mitigating climate change, grasslands are particularly suitable to many pollinators and therefore sustain biodiversity and food production in both extensive and intensive agricultural systems, and pastures. Third, this spatial information is essential to make management plans, effectively study the environment and conserve natural spaces.

Now this is where the magic happens. By projecting climate and land-cover changes over time, it becomes possible to evaluate the vulnerability of species, to model carbon storage or pollination in the future, or to evaluate the relevance and feasibility of a management plan. But while climatic projections are widely available and modelled for many scenarios depending on our ability and velocity in reducing greenhouse gases’ emissions, projections of land-cover changes are much more complicated to model. But why ?

In many places around the world, land-cover is highly intertwined with human management and only some habitats in remote places are purely natural. This means that without human intervention and disturbance, the environment would be very different. Therefore, in managed landscapes, modelling and projecting land-use changes does not really depend on physical properties like climate, but mostly on the environmental policies that will be implemented in the future. This is much more complicated to predict, if ever possible, because it relies on the vision and ambition of stakeholders, as well as the challenges and pressures that will arise in the coming years. So how to get around this problem ?

George EP Box and Norman Draper wrote that “essentially, all models are wrong, but some are useful” (1987), and to be useful, a model has to take into account the range of possibilities that is likely to occur, for example by considering several scenarios of plausible futures. In order to create relevant, reality-anchored scenarios from which land-cover changes will be modelled and projected, a workshop has been organised with the stakeholders from all regions of the Tagus river catchment. Based on existing literature and the discussions with the stakeholders, three scenarios were retained for the end of the century.

The first one is business-as-usual involving a significant loss of biodiversity and quality of habitats due to an increased intensive agriculture and urbanization. Nature’s protection laws are dropped and only a few remaining strictly protected areas are kept. The second one is much more optimistic and reflects what would happen in a best case scenario where sustainability and self-sufficiency would be the main objectives. Here, biodiversity is bouncing back, conservation areas become much more protected, urban areas are not expanding and intensive agriculture is being replaced by extensive practices. Finally, the last scenario is a neutral/realistic forecast in which intensive uses and urbanization increase, but so do the natural and protected areas.

Figure 1 - Current land-cover for central Tagus river catchment near Talavera de la Reina (Spain)

Now, the distribution of future land-cover can be projected, based on:

1) the scenarios developed by the stakeholders

2) realistic plausible changes for each category of land-cover in the coming decades based on the literature

3) observed changes from past to present times

Notice the current situation (Fig. 1) and the visual impacts these changes of land-cover can have on the area. In the business-as-usual scenario (Fig. 2), we see a large predicted increase of intensive agriculture (in yellow) and even urban areas (in pink) around transportation roads at the expense of natural habitats such as forests. On the contrary, in the best case scenario, intensive agriculture is being replaced by extensive practices (in bright green), while forested areas are kept and grasslands expand.

 Figure 2 - Projected land-cover for 2080 in best case scenario (left) and business-as-usual scenario (right) on the same location

These projections are the elementary bricks that can be used for further analysis. Exploring the literature and co-developing relevant scenarios anchored in reality with the stakeholders participate in identifying the range of plausible evolutions of land-cover in the Tagus river catchment. These results, in turn, will help develop better management practices for the conservation of biodiversity.

References

George EP. Box & Norman Draper (1987), Empirical Model-Building and Response Surfaces, p.424, 1987

IPBES (2019), Global assessment report of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, Brondízio, E. S., Settele, J., Díaz, S., Ngo, H. T. (eds). IPBES secretariat, Bonn, Germany. 1144 pages. ISBN: 978-3-947851-20-1

Sage, R. F. (2020). Global change biology: a primer. Global Change Biology, 26(1), 3-30.