tions. This was, for example, a key ingredient of the U.S. GLOBEC program2 in which short-term process studies were embedded within the framework of a monitoring program.
Monitoring is a synergistic component in modeling and hypothesis development. It provides datasets necessary for the evaluation and development of models and/or suggests investigations needed to improve model parameterizations and/or processes. Models provide an integrated approach to understanding system behavior and can be used to modify the monitoring program as necessary. Models also augment monitoring efforts by suggesting how unsampled system components may be evolving. Monitoring and model results both contribute to the construction of hypotheses on how the system or parts of it operate.
Much of our recognition and understanding of the dramatic changes occurring in the Arctic has emerged from long-term observations. For example, routine measurements revealed the dramatic warming of the Arctic atmosphere and the accelerating decline in sea ice; both are consistent with some of the earliest model predictions of climate response to greenhouse gas warming (Manabe and Stouffer, 1980). Another example is the systematic approach adopted by the Arctic and Bonanza Creek Long-Term Ecological Research (LTER)3 programs conducted in the tundra and boreal forest biomes of Alaska, respectively. Although independently initiated, these LTERs are established along a latitudinal and ecological gradient and each attempts to understand the resiliency and vulnerability of the respective biome to a warming climate. Both LTERs have been in existence for at least 25 years and involve myriad interdisciplinary process studies and modeling activities. Although different investigators are involved in each, there are consistent efforts to compare and contrast the results across biomes.
One important finding from the integration of plot-scale long-term studies of vegetation dynamics, fire cycles, and their links to climate in the Bonanza Creek LTER (Van Cleve and Vierech, 1981; Van Cleve et al., 1983) with broader-scale measurements of a series of wildfire-disturbed boreal forests of interior Alaska is the likely shift in some Alaskan boreal forests from a spruce-dominated to a broadleaf-dominated landscape due to increased burn severity (Figure 4.2). This transition to more high-severity wildfires is occurring in conjunction with thawing of permafrost and the decomposition of previously frozen organic carbon in boreal forest soils. Through large-scale manipulation experiments at the Arctic LTER at Toolik Lake, researchers have found that response to heating soil, shading, or altering soil moisture is slow, with responses delayed until 9 or 10 years post initiation of the treatment (Hobbie and Kling, 2014).