An assertion has been advanced that drought is defined by its impacts. This stems from a definition of drought that considers amount of available water in relation to needs for that water. Where people and their activities are present, economic and other measures can be employed to portray costs and losses resulting from abnormal dryness. These can in turn be utilized to determine the magnitude or severity of drought. In the ideal case, metrics that can be computed in a consistent way over many decades can be analyzed to determine the historical context and relative rarity of a drought of interest. For human systems, temporal heterogeneity over long periods is elusive, as economic sectors rise and fall through time, as metrics of such activities vary by sector, and as population and economic base (usually) grows.

Natural systems can be considered to have intrinsic value and utilitarian value. Live forests filter air, provide oxygen, serve as wildlife habitat, act as erosion buffers and hydrologic filters, and provide aesthetic enjoyment. These all have value, even if hard to quantify. Moisture-stressed forests, dying or dead, are less able to provide these outcomes. In addition they foster pest infestations, increase fire likelihood, attract fewer human visitors, and have utilitarian impacts to human systems such as loss of harvestable trees. These impacts can be quantified to some degree. Wilderness areas, protected from most large-scale direct human impact, offer a good example. They cannot be harvested, so drought induced impacts (e.g., insect or fire mortality) that utilize value as timber, for example, are essentially only hypothetical. Furthermore, in forests that evolved with episodic drought, dry episodes can be viewed as having both positive and negative consequences. The drought of 1988 dried Yellowstone to such an extent that huge fires could be maintained. Individual trees were killed but the forest survived, and a new generation is now even thriving. Was the net long-term impact of that drought negative or positive?

The U.S. Drought Monitor defines its "D-levels" in terms of return period or percentile. These levels can be computed from geophysical information such as accumulated precipitation or streamflow. Complications in this determination that arise from timescale-dependent and simultaneously present differences in such percentiles have not been adequately resolved, although this issue has been acknowledged in serving as the motivation for the Standardized Precipitation Index. The physical conditions experienced in the Yellowstone example were indeed rare, and certainly deserved a large D-value, even with the absence of significant direct human influences. These and related issues will be explored further in this presentation.