The Community Land Model (CLM5) is widely used by the Earth System Modeling research community to study many aspects of the role of land in climate and weather. In particular, the omodel is frequently used to understand and predict global and regional land carbon stock trajectories, water state trends, and carbon-water interactions and water use efficiency trends. Recent work has demonstrated high uncertainty due to forcing, structural, and parametric uncertainties. Prior efforts to assess CLM parametric uncertainty have been hampered by computational constraints or code limitations, necessarily limited to selected parameters related to specific processes. Here, we present a new community effort to conduct a comprehensive tiered exploration of parameter sensitivity and uncertainty; the CLM5 Parameter Perturbation Ensemble project (CLM5PPE). We have identified 200+ model parameters across processes that control energy, water, carbon, and nitrogen interactions. Phase 1 of the CLM5PPE involves one-at-a-time high/low parameter perturbations for all 200+ parameters on a sparse grid (~250 grid cells) that reasonably captures the main features of global higher-resolution simulations. Each simulation is checked for reasonableness (e.g., vegetation survivability rates). Each parameter perturbation is also run with environmental perturbations (CO2, climate, N-deposition) that span historical and projected values. A set of 50 parameters are selected for further evaluation with the criteria for selection based on their importance in determining the mean, variability, and responses to environmental perturbations for a range of key land climate variables. Phase 2 uses these parameters to run a Latin hypercube sparse-grid 2500-member perturbed parameter ensemble, again repeated for each environmental perturbation. In Phase 3, ~200 best performing parameter sets will be used to run an ensemble of historical and projection period 2o resolution simulations to provide a realistic and comprehensive assessment of parametric uncertainty. All data output from this project as well as the scripting infrastructure to automate parameter perturbations, generate large ensembles, and assess model performance will also be made available to facilitate further parameter exploration of this and future versions of CLM.