In the first phase of PELD, little interaction between sites occurred effectively. Integration difficulties are partly due to differences between ecosystems, but mainly to the lack of standardization between research areas. The lack of standardization makes comparisons between sites difficult, since even the same processes, but studied at different scales, are not comparable.
In a preliminary standardization effort, 4 PELD sites (sites 1 - Tropical Humid Forest, 2 - South Pantanal, 11 - North Pantanal and 12 - Caatinga) established a standardized experimental design, based on the RAPELD system (Magnusson et al. 2005) . This system consists of a grid of trails measuring at least 5 x 5 km, with permanent plots, demarcated in a contour line, distributed every 1 km, and a complementary series of "special" plots designed to cover as evenly as possible environments with more restricted distribution (for example, watercourses, in forests, or hilltops in flooded landscapes). In this phase, studies on the RAPELD grids focused mainly on determining the basic environmental characteristics (topography, texture and fertility of the soil, vegetation structure, physical and chemical characteristics of the water, morphological characteristics of the watercourses) of the permanent plots and in to map the distributions of individual organisms and communities in relation to the characteristics of the environment, although some population and community dynamics studies have also been developed.
With this environmental information base, and with the initial inventories of populations and communities conducted, the strategy for phase II will be to implement the temporal monitoring of selected populations and communities in all sites with a RAPELD grid that are interested in collaboration, always using comparable methods between sites. It is not necessarily intended to monitor the same species at each site, but species from the same biological groups, which can serve as indicators of direct anthropic changes and climate change. Especially to monitor the effects of climate change, it will be necessary to monitor species typical of each ecosystem, which are generally not the same between ecosystems.
The idea is to be able to compare the intensity of the impacts of changes on a global scale across ecosystems, using information on a comparable scale, on specific indicators. The exchange of information between the RAPELD grids will be sought to standardize methodologies and approaches, and to exchange experiences.
The studies on population and community dynamics, and vegetation dynamics implemented in phase I of PELD Manaus formed the basis of discussions on the monitoring program of the Protected Areas of the Amazon program (ARPA, IBAMA/WWF). The RAPELD design was adopted by ARPA, as well as some sampling protocols. The managers of ARPA protected areas are very interested in the databases generated by the sites that use RAPELD, as these constitute bases for comparing the dynamics observed in their areas. For phase II of ARPA, training of local agents is planned to monitor some of the same components included in this proposal. Part of this training will be carried out at Reserva Ducke, by researchers involved in this proposal.
Likewise, the RAPELD design was adopted by IBAMA Environmental Licensing, and therefore, comparable databases in undisturbed locations are important for this sector, so that they can judge the results of monitoring the impacted areas. PELD site I researchers have been involved in training IBAMA's environmental analysts to build an integrated environmental monitoring system. In this way, the expectation is that several areas of the Amazon will be monitored with similar design and protocols in the coming years.