Explain Future challenges of connecting models and data?
Since biological systems are complex, models of biological systems are also complex, and matching models and data is a challenge. Primary to fitting models to data is separating the "deterministic skeleton" explaining the interactions among components of the system from the stochastic variables that collision the system. The dynamics of deterministic skeletons may be difficult in the nonexistence of stochasticity, and stochasticity might drive dynamics beyond those showed by the deterministic skeleton (Rand and Wilson 1991, Henson et al. 2001). Additionally, there may be multiple sources of stochasticity acting on various components of the system. One way to these problems is developing "semi-mechanistic" models (Ellner et al. 1998) that entail those components of a biological system which we know for specific, while absorbing less famous components into more general, stochastic terms.
In spite of recent advances, developing methods to disentangle the deterministic and stochastic components of biological systems is yet in its infancy. Data from environmental systems are fraught with error. Error takes part from our inability to see major components of the system, or our inability to measure the system precisely. It is often essential to model biological systems by separating the mechanistic description of the system (process model) from the way in that we observe it (observation model) (Harvey 1989, Schnute 1994, de Valpine and Hastings 2002). When the construction of such state-space models is here an active area of research in statistics, little of this research has permeated environmental and integrative sciences. One of the biggest challenges come across modern quantitative environmental and integrative biology is connecting the large, dispersed, variable quality data sets along with the existing body of ecological theory. In short order, remote sensing and microsensor data will add vast quantities of spatially referenced data to the milieu. How can these data be employed and related with the body of ecological theory?