José Manuel Cascalho

Automated discovery of relationships, models and principles in ecology

Ecological systems are the quintessential complex systems, involving numerous high-order interactions and non-linear relationships. The most commonly used statistical modelling techniques can hardly reflect the complexity of ecological patterns and processes. Finding hidden relationships in complex data is now possible through the use of massive computational power, particularly by means of Artificial Intelligence methods, such as evolutionary computation. Here we use symbolic regression (SR), which searches for both the formal structure of equations and the fitting parameters simultaneously, hence providing the required flexibility to characterize complex ecological systems. First, we demonstrate how SR can deal with complex datasets for: 1) modelling species richness; and 2) modelling species spatial distributions. Second, we illustrate how SR can be used to find general models in ecology, by using it to: 3) develop species richness estimators; and 4) develop the species-area relationship and the general dynamic model of oceanic island biogeography. All the examples suggest that evolving free-form equations purely from data, often without prior human inference or hypotheses, may represent a very powerful tool for ecologists and biogeographers to become aware of hidden relationships and suggest general theoretical models and principles.

Towards an Agent Based Modeling: The Prediction and Prevention of the Spread of the Drywood Termite Cryptotermes brevis

We present initial efforts made to model the spread of the drywood termite in Angra do Heroismo, Azores, using an agent based modeling approach. First we describe how a simple Cellular Automata (CA) model was created in Netlogo to simulate the spread of the species based on simple assumptions concerning the ecology of the species. A second step was taken by increasing the complexity of the initial CA approach, adding new specific characteristics to each cell, based again on ecology of the species and its behavior towards the environment. Finally, we add agents to the model in order to simulate the human intervention in fighting back the pest. This new model has become a two-level Agent-Based model. We also evaluated the costs of this intervention. These efforts were supported by field research which allowed a continuous cross-checking of the results obtained in the model with the field data.

Assessing the landscape functional connectivity using movement maps: a case study with endemic Azorean insects

There is a vast body of literature aiming to predict, for a large number of taxa, the spatial distribution of suitable areas given the expected future changes of climatic conditions. However, such studies often overlook the role of landscape functional connectivity. This is particularly relevant for species with low vagility, as ground-dwelling insects, inhabiting areas with high human pressure due to habitat destruction and fragmentation, namely in the islands. In this study, we developed an individual-based model (IBM) that simulates individual movement according to landscape resistance and mortality probability, in order to derive the landscape movement map, and applied it to five endemic ground-dwelling insects of Terceira Island (Azores). We then confronted the movement maps of each species against the species distribution models previously developed for both current and future climatic conditions, quantifying the amount of important movement areas that are enclosed by the distribution polygons. We further sought to identify where habitat restoration would increase the overall connectivity among large habitat patches. Our results showed that, for both timeframes, the distribution models enclosed small amounts of areas predicted to be important for animal movement. Additionally, we predicted strong reductions (up to 94%) of these important areas for functional connectivity. We also identified areas in-between native forest of primary importance for restoration that may significantly increase the probability of persistence of our model species. We anticipate that this study will be useful to both conservation planners and ecologists seeking to understand species movement and dispersal both is islands and elsewhere.

Agent-Based Modelling for a Resource Management Problem in a Role-Playing Game

In this paper we present a prototype of a model created in the context of a resource management problem in Gaza, Mozambique. This model is part of a participatory approach to deal with a conflict of water supply. Farmers and cattle producers are added to a stylized environment and a conflict is modelled when cattle needs to access water and destroy farmers’ harvest. To address the different behaviours of farmers and cattle producers, a BDI architecture is used to support the conflict simulation using a simple argument-based negotiation between proactive agents. This model is intended to be used as a support to a Role Playing Game (RPG) in the context of an interactive design assembled under the Netlogo tool environment.