Anomalous atmospheric circulation features are a prime source of drought over the US and can
often be associated with remote SST influence, often modulated by other factors. Once drought
is underway, drying soil and withering vegetation may exacerbate drought locally and regionally
by reinforcing conditions inimical to precipitation. The proposed project will investigate the
holistic drivers and responses to drought over the US in an Earth system context using
observations, reanalyses, subseasonal forecasts and global model sensitivity studies. Given
that probabilistic predictability is derived from both land and ocean components on subseasonal
to seasonal time scales, there is great potential to harvest more forecast skill in a complete
approach. Three questions are posed: (1) What are the combined roles of ocean and land in the
initiation and maintenance of drought over the US as suggested by observationally based data
sets? (2) Do forecast models employed on time scales relevant for drought onset and demise
(subseasonal to seasonal) replicate the relationships indicated in observational data? (3) Can the
interplay between remote ocean and near-field land anomalies as drivers of drought be
diagnosed in sensitivity studies using the latest global forecast model?
The proposed research has three objectives: (1) Characterize the roles of ocean and land in
drought by examining the coupled processes each have with the atmosphere in a systems
approach using established metrics, updated with techniques from information theory that
minimize arbitrary assumptions; (2) Diagnose operational and research models that currently
provide subseasonal forecasts, portraying their coupled ocean-land-atmosphere performance,
and potentially attributing drought forecast skill to specific model behaviors; (3) Specifically
apply and diagnose the behavior of the Unified Forecast System (UFS) in the area of drought
prediction, including sensitivity studies to isolate drivers.
Metrics of surface-atmosphere interaction will be used to determine the spatial and temporal
variation of process chains that link surface anomalies to drought. Methods from information
theory will provide novel extensions to convention linear/Gaussian based statistics, giving
nonparametric estimates of active process networks in the Earth system. Observational data
will be used to provide a model-free estimate of drought driver relationships and responses,
which will be compared to reanalysis-based estimates to provide an independent verification
data set for forecast model evaluation. Existing retrospective subseasonal forecasts will then be
evaluated for process fidelity, including lagged responses important for operational forecasting.
The Unified Forecast System (UFS) will also be evaluated and used for sensitivity studies to
explore specific US drought cases.
The research will address all three drought research priority areas by (1) identifying surface-
atmosphere interactions and their related processes that lead to drought; (2) identifying key parameters,
applying methodologies and pathways to use coupling metrics that can contribute
to the capacity of NIDIS to identify situations of elevated drought predictability and risk; (3)
Associating processes and feedbacks between land, ocean and atmosphere that contribute to
drought predictability, prediction and warning.