Table 1.

Four categories of uncertainties encountered in effects analyses and implementation of adaptive management (derived from Buneau et al. 2015)

Environmental variability	The probability of many environmental phenomena, including episodic events such as wildfires and earthquakes, near-term weather extremes, and future climate, is uncertain. Drivers of habitat extent and quality, such as flow levels in river systems, annual variability in the phenologies of growing seasons, the distribution of temperature maxima and precipitation, and the presence and abundance of predators and prey, are prime determinants of the distribution and population dynamics of species. Yet these sources of uncertainty are largely irreducible. Advances in modeling and expanded time-series data sets can lead to better estimates of the likely distribution of future conditions and target species responses
Structural uncertainty	Although the fundamental relations between physical conditions at landscape and smaller extents, habitat quantity and quality, and reproductive success sometimes can be inferred from available data, uncertainties inevitably remain concerning the functional form of some relations. What aspects of landscape condition vary in what spatial and temporal patterns to affect habitat extent and quality, how does habitat condition affect local population and metapopulation dynamics. Structural uncertainty can be reduced through research, monitoring, and improvements to models
Parametric uncertainty	Even where the structure of ecological relationships is well known, uncertainty can remain as to the strength of those relationships. For example, what amount of habitat for an imperiled shorebird is available at a given river stage, what minimal abundance of a rare plant is required to support its pollinators, and what salinity level is tolerated by an estuarine fish at each life stage. As with structural uncertainty, those uncertainties can be reduced through research and monitoring and incorporated into models; however, varying over time and by location they can resist resolution
Observation uncertainty	Neither estimates of population size and reproduction, nor habitat structure and composition can be fully accurate. Degrees of error and direction of bias can vary with species characteristics, habitat attributes, and level of sampling effort, thus differ across both space and time. Rigorous design and level of effort in a monitoring program can reduce observation error and, in some designs, estimate the error in targeted surveys, which allows for more accuracy the resulting information