Modelling the dynamics of Mediterranean forests at large scales is a challenge. But scientists are setting up simulation models to help us better understand future directions of these ecosystems and how to better manage them.

Throughout history, Mediterranean forests have been highly modified by human actions and their current situation is the result of a constant impact which has altered ecosystems in multiple ways. In the early 1900s, forested landscapes in Spain were considered being in a deplorable state after centuries of deforestation, agriculture and silvicultural land use. To face this situation, an intense reforestation policy was put into place, converting almost 4 million hectares of barren terrain to (mostly) pine plantations. The aim of this reforestation policy was to halt the high erosion rates resulting from the abandonment of agricultural lands and eventually transform these areas to forest ecosystems. More specifically, the reforestation plan aimed to establish pines only as an intermediate stage before more diverse systems could emerge. However, over the years appropriate management towards this goal has been generally missing. As a result, current pine plantations in Spain often show high stand densities and low levels of biodiversity. This situation is not ideal because in these very dense stands only a small portion of available light is able to reach the understory, compromising natural regeneration and future stability of these ecosystems. In areas where light conditions allow regeneration and where seeds of other trees species disperse from nearby stands, a mixed forest is timidly appearing.

In a future characterized by a rapidly changing climate change and growth stagnation of these pine plantations, the future of forest ecosystems in Spain and other Mediterranean areas is uncertain. Longer dry periods, higher incidence of forest pests and increasing intraspecific competition increase the risks of forest decay and massive forest fires. Forest management can probably help to reduce these risks, but anticipating the effects of silvicultural treatments is tricky since we are facing unprecedented climate conditions.

Luckily for us, nowadays ecologists are equipped with a variety of tools to better understand forests functioning and predict future changes. Some of these tools are forest landscape models . These models can be used to simulate multiple ecological processes occurring simultaneously in forests at different temporal and spatial scales. They are great ways to perform experiments and get a deeper insight into forests functioning.

Forest landscape models simulate the landscape as a set of cells in a grid. Each cell has a series of attributes such as presence of a species and its age (or other attributes). A set of functions defines how species grow, reproduce, disperse or die, therefore resulting in a program which can simulate forest succession under different experimental conditions. Multiple forest landscape models exist with a wide range in complexity and a gradient from more empirical to more mechanistic approaches (see this article where scientists compared several of them on the same landscape).

>> Read more: Are ash trees doomed?

Can we model Mediterranean species growth?

Mediterranean summers are dry and hot, and so species living under this climate limit their growth during the summer season in order avoid too high evapotranspiration rates. Thus, Mediterranean species often show two peaks of growth through the year, typically during spring and fall. Due to this bimodal pattern, correctly capturing the growth of Mediterranean species using dynamic models is often a challenge.

In order to appropriately simulate forest succession in Mediterranean pine plantations, a forest landscape model able to simulate such bimodal growth is needed. This can be achieved by using process-based approaches (models that explicitly simulate photosynthesis and respiration processes) with a temporal resolution that capture inter-annual growth patterns (i.e., day or month instead of year). An example of such approach is represented by PnET-Succession – an extension of the well-known forest landscape model LANDIS-II – that simulates tree growth in a mechanistic way and at a monthly scale. We calibrated several Mediterranean tree species in the model and found that PnET-Succession reproduces reasonably well the expected growth pattern for Pinus and Quercus spp. in two typical forest types in our area (maritime and Aleppo pine with holm oak; Scots and black pine with Pyrenean oak).

>> Read more: Giant tree hunting and old-growth forest conservation in Vancouver Island

At a larger scale…

In order to simulate forest succession in Mediterranean pine plantations, we have selected a representative landscape in southeastern Spain. Sierra Nevada, Sierra de Baza and Sierra de Huétor National and Natural Parks are great study sites to investigate the fate of Mediterranean pine plantations. This area covers an altitude range between 500 m and 3478 m a.s.l. therefore resulting in a wide range of environmental conditions in terms of temperature and precipitation. Because of its altitudinal gradient, this area shows a great representation of Mediterranean ecosystems and is home of multitude of endemic species. The main tree species present in this area are pines (Pinus sylvestris, P. nigra, P. pinaster and P. halepensis) and oaks (Quercus ilex , Q. pyrenaica and Q. faginea).

Additionally to parameterization of tree species growth, setting up a complex simulation model such as LANDIS-II in a new landscape of application is very time consuming. It requires collecting all the necessary data (e.g., climate, soil, forest inventory) that need to be prepared in the adequate format for the model. Ecological models all agree that model initialization and parameterization phases are those requiring most of the efforts in terms of time. For example, one of the crucial inputs for most forest landscape models is a vegetation map describing initial forest conditions at the start of the simulation experiment (e.g. presence of tree species across the landscape). To generate such map, we used the Spanish National Forest Inventory as well as two vegetation cover maps. The integration of multiple information sources to generate maps of vegetation is a time demanding process but often not very well described in literature studies and usually buried within methodological details. However, a clearly described process to generate vegetation cover maps for landscape model can be quite interesting and could encourage others to start using forest landscape models. This is the reason why we documented this process in a step-by-step approach in this article that we recently published in the journal Frontiers in Ecology and Evolution.

So far, our efforts have been targeted on getting the landscape ready and parameterized for applications and we are now working on the design of management scenarios which can help us explore how silvicultural treatments can modulate the effects of climate change in our systems. We aim to simulate the future of our pine plantations, and this way being able to assist stakeholders in decision-making processes in the context of forest management under global environmental changes. Stay tuned!

Read here the paper open access:

Suárez-Muñoz M, Mina M, Salazar PC, Navarro-Cerrillo RM, Quero JL, Bonet-García FJ. 2021. A step- by-step guide to initialize and calibrate landscape models: a case study in the Mediterranean mountains. Frontiers in Ecology and Evolution. 9, 209.

Main photo: Mixed conifers-oaks forest in Sierra de Baza, Spain. Originally dense pine plantation has developed into an open mixed forest (photo by M. Suarez-Muñoz)

Authors of the post

María Suárez-Muñoz is a PhD candidate at the University of Granada and environmental scientist at the Instituto Interuniversitario del Sistema Tierra de Andalucía, Spain. She has a background in environmental science from the Pablo de Olavide University in Sevilla and a master from the University of Amsterdam. For her PhD project she is investigating the dynamics of pine plantations and Mediterranean ecosystems using landscape dynamic models.

Marco Mina is a forest ecologist, postdoctoral fellow at the Centre for Forest Research at the Université du Quebéc à Montreal, Canada. His research interests include the dynamics of temperate forest ecosystems and anticipating the impacts of global and climate change using mathematical models of forest dynamics to explore interactions of trees with their changing environment and to investigate management strategies to enhance forest resistance and resilience to future challenges.