Towards a virtual observatory for ecosystem services and poverty alleviation

For the indigenous communities of the Pacaya-Samira National Reserve in the Peruvian Amazon, turtle farming is a successful survival strategy. The practice keeps the natural animal population numbers up and provides a necessary source of food and income. However, the success of the activity is highly dependent on information about the erratic river level, which may flood nesting beaches at crucial times. In the Yasuni national park in the Ecuadorian Amazon, bush meat hunting regions are threatened by encroaching deforestation. At the same time in the Andean headwaters of Ecuador, Peru and Bolivia, the availability and quality of irrigation water depends strongly on mountain wetland management, and is potentially threatened by global climate change. These are striking examples of many situations where the livelihoods of local communities depend on crucial ecosystem services. However, a sustainable management of these services is only possible using an advanced integration of climatic, hydrological, and ecological data. Current approaches to integrate such data have largely failed for a variety of reasons. In Pacaya Samira, little local data are available resulting in very large uncertainties in the model predictions. In the Andean highlands, local politicians and managers have difficulties interpreting model simulations and design proper land management schemes. Finally, both systems can benefit strongly from the incorporation of local expert knowledge to reduce model uncertainties. Recently, many methodologies for data integration and user interaction have been developed. They are known under the common umbrella of a 'virtual observatory' (VO). The ultimate goal of a VO is a perfect integration of data, models and users. Worldwide, many coordinated activities are ongoing to make this integration a reality. However, far less attention has been paid to the question of how these developments can benefit environmental services management and poverty alleviation. This project will design and implement an environmental prediction system for the above mentioned case studies, using existing virtual observatory tools. In a next step, we will develop, implement and evaluate tools to improve the value of these systems in the specific conditions of poverty alleviation, i.e., (1) Improved communication of simulations. This action will particularly focus on the visualisation of modelling results and their uncertainties; (2) Assessing the value of collected data. In a data sparse and resources constrained environment, an optimal design of new data collection strategies is essential. Here we will develop methods to simulate the value of different data on the model predictions; (3) Integrating local managers' knowledge and practice in modelling systems. This module deals with the development of a user interface to evaluate models, identify model failures and reject models. Heavily relying on public domain software, open standards and existing VO efforts, we will develop a platform for interdisciplinary, cost-efficient and highly tailored environmental data analysis and simulation. This platform will be available immediately for the selected case studies, thus enabling direct poverty alleviation action benefiting an estimated 15000 local inhabitants. Close collaboration with local stakeholders and integration in existing initiatives ensures a quick adoption of the platform. For instance, the InfoAndina website of project partner CONDESAN which will be integrated, has more than 1600 registered users. At the same time, the project will generate novel scientific insights in model simulation, communication and improvement in a developing context. The involvement of the PI in the global Virtual Observatory community will ensure that the research results will optimally benefit ongoing research in this area.