Contributing factors for drought in United States forest ecosystems under projected future climates and their uncertainty

•Complex changes to drought through multiple processes hidden by summary metrics.•Some drought-related climate changes like dry-spell length or early snowmelt relatively certain.•Others related to seasonal and annual precipitation are substantially uncertain.•Increasing vapor pressure deficit (VPD) and temperature are relatively certain.•Focus of VPD consequences should be on physiology, not changing water balance. Observations of increasing global forest die-off related to drought are leading to more questions about potential increases in drought occurrence, severity, and ecological consequence in the future. Dry soils and warm temperatures interact to affect trees during drought; so understanding shifting risks requires some understanding of changes in both temperature and precipitation. Unfortunately, strong precipitation uncertainties in climate models yield substantial uncertainty in projections of drought occurrence. We argue that disambiguation of drought effects into temperature and precipitation-mediated processes can alleviate some of the implied uncertainty. In particular, the disambiguation can clarify geographic diversity in forest sensitivity to multifarious drivers of drought and mortality, making more specific use of geographically diverse climate projections. Such a framework may provide forest managers with an easier heuristic in discerning geographically diverse adaptation options. Warming temperatures in the future mean three things with respect to drought in forests: (1) droughts, typically already unusually hot periods, will become hotter, (2) the drying capacity of the air, measured as the vapor pressure deficit (VPD) will become greater, and (3) a smaller fraction of precipitation will fall as snow. More hot-temperature extremes will be more stressful in a direct way to living tissue, and greater VPD will increase pressure gradients within trees, exacerbating the risk of hydraulic failure. Reduced storage in snowpacks reduces summer water availability in some places. Warmer temperatures do not directly cause drier soils, however. In a hydrologic sense, warmer temperatures do little to cause “drought” as defined by water balances. Instead, much of the future additional longwave energy flux is expected to cause warming rather than evaporating water. Precipitation variations, in contrast, affect water balances and moisture availability directly; so uncertainties in future precipitation generate uncertainty in drought occurrence and sev