To address the identified next generation Benchmarking Challenges and Priorities, certain Enabling Capabilities are needed. New model development focused on improving process representations is required, and additional model variables should be saved for comparison with data. A new Land Model Testbed (LMT) capability employing community benchmarks and supporting UQ frameworks would enable more rapid model development and verification, particularly for major ESM frameworks like ACME and CESM.Additional field measurements and monitoring activities, as well as perturbation experiments and lab studies, could provide valuable observational data for constraining models. High priority measurement needs for developing benchmarks and improving ESMs include the following: › Long-term energy, carbon, and water flux measurements at AmeriFlux and FLUXNET sites with standardized instrumentation and methods, and additional frequent or continuous ancillary in situ measurements of soil moisture, sap flow, tree height and diameter, litterfall, and soil nutrients › High latitude and tundra soil core measurements of carbon and nutrient distributions, including isotopes and ice/water content, and observations of vegetation growth and expansion of woody vegetation › Characterization of tropical ecosystem traits and canopy structure and chemistry; observations of tropical ecosystem responses to drought, increased temperatures, and elevated atmospheric CO 2 ; and measurements of nutrient cycling and hydrology in tropical forests, focusing on land-atmosphere interactions › Remote sensing algorithms and processing infrastructure for generating data products useful for large-scale ecosystem characterization and monitoring, scaling up in situ measurements, and informing future measurement site selection.Improved observational data archives (e.g., DOE Atmospheric Radiation Measurement (ARM) Climate Research Facility and Environmental System Science (ESS) archives, NASA Distributed Active Archive Centers (DAACs)) and repositories (e.g., Obs4MIPs) are needed that offer data discovery, server-side analysis, and advanced distribution capabilities. Finally, new computational resources and cyber infrastructure will be required to realize the promise of new benchmarking capabilities. This infrastructure needs to offer a balance between pure compute capacity (high core count) and throughput (e.g., cache size, memory size and bandwidth, and input/output bandwidth) to support in situ analysis and benchmarking, growing observational data sets, and multi-model comparisons.
Conclusions and Next StepsThe 2016 ILAMB Workshop was successful in bringing together the international community to identify scientific challenges and priorities for future research. The workshop demonstrated broad interest on the part of a vibrant community of scientists spanning many disciplines that are committed to reducing barriers for information flow between the measurement and modeling communities.To effectively address the individual processes and cross-cutting themes discus...