Long-term ecological research
The life span of ecosystems is thousands of years. As a consequence, many ecological questions on ecosystem evolution and forecasting cannot be addressed with short-term observations or experiments. To face this problem a Long Term Ecological Research network (LTER) was established in 1980. LTER joins research projects addressing ecological phenomena over long periods. In this frame more than 30 years ago we started collecting data on the microbial food chain and related variables with high temporal (monthly, fortnightly) and spatial frequency along the water column of Lake Maggiore. The time series of data thus gathered significantly advanced our understanding of the long-term dynamics of the microbial communities and of this lake.
It has been possible to ascertain the oligotrophication process of the lake and the consequent changes in its organic carbon content (Bertoni R. and C. Callieri 1992) from the qualitative point of view (Bertoni R., C. Callieri, G. Morabito, M.L. Pinolini and A. Pugnetti 1997). The horizontal heterogeneity of organic carbon and microbial communities was also documented, which is useful to set the confidence on data often coming from a single sampling station (Bertoni R., R. Piscia and C. Callieri 2004).
The long-term data set made also possible to show how the physical constrains are driving planktonic Bacteria and Archaea distribution (Bertoni R., W. Ambrosetti and C. Callieri 2007)
The coupling of bacteria and organic carbon in Lake Maggiore was also highlighted along long-term trends (Bertoni R., C. Callieri, G. Corno, S. Rasconi, E. Caravati and M. Contesini 2010)
Finally, based on a 30 year-long data series a non-deterministic approach to forecasting the trophic evolution of Lake Maggiore was tested. The use of genetic programming made it possible to include, further to microbial and organic carbon variables, also hydrochemical and meteorological variables to generate a robust forecasting model tested on a 5 years set of data (Bertoni R., M. Bertoni, G. Morabito, M. Rogora, and C. Callieri 2016).
* in the figure: Comparison of Genetic Programming and MLR forecast of TOC concentration