Kenneth W. Harrison

A philosophical basis for hydrological uncertainty

ABSTRACT Uncertainty is an epistemological concept in the sense that any meaningful understanding of uncertainty requires a theory of knowledge. Therefore, uncertainty resulting from scientific endeavors can only be properly understood in the context of a well-defined philosophy of science. Our main message here is that much of the discussion about uncertainty in hydrology has lacked grounding in these foundational concepts, and has resulted in a controversy that is largely the product of logical errors rather than true (axiomatic) disagreement. As an example, we explore the current debate about the appropriate role of probability theory for hydrological uncertainty quantification. Our main messages are: (1) apparent (and/or claimed) limitations of probability theory are not actually consequences of that theory, but rather of deeper underlying epistemological (and ontological) issues; (2) questions about the appropriateness of probability theory are only meaningful if posed as questions about our preferred philosophy of science; and (3) questions about uncertainty may often be better posed as questions about available information and information use efficiency. Our purpose here is to discuss how hydrologists might ask more meaningful questions about uncertainty.

Impact of Land Model Calibration on Coupled Land–Atmosphere Prediction

AbstractLand–atmosphere (LA) interactions play a critical role in determining the diurnal evolution of both planetary boundary layer (PBL) and land surface heat and moisture budgets, as well as controlling feedbacks with clouds and precipitation that lead to the persistence of dry and wet regimes. In this study, the authors examine the impact of improved specification of land surface states, anomalies, and fluxes on coupled Weather Research and Forecasting Model (WRF) forecasts during the summers of extreme dry (2006) and wet (2007) land surface conditions in the U.S. southern Great Plains. The improved land initialization and surface flux parameterizations are obtained through calibration of the Noah land surface model using the new optimization and uncertainty estimation subsystems in NASAs Land Information System (LIS-OPT/LIS-UE). The impact of the calibration on the 1) spinup of the land surface used as initial conditions and 2) the simulated heat and moisture states and fluxes of the coupled WRF sim...

Quantifying Uncertainties in Land-Surface Microwave Emissivity Retrievals

Uncertainties in the retrievals of microwave land-surface emissivities are quantified over two types of land surfaces: desert and tropical rainforest. Retrievals from satellite-based microwave imagers, including the Special Sensor Microwave Imager, the Tropical Rainfall Measuring Mission Microwave Imager, and the Advanced Microwave Scanning Radiometer for Earth Observing System, are studied. Our results show that there are considerable differences between the retrievals from different sensors and from different groups over these two land-surface types. In addition, the mean emissivity values show different spectral behavior across the frequencies. With the true emissivity assumed largely constant over both of the two sites throughout the study period, the differences are largely attributed to the systematic and random errors in the retrievals. Generally, these retrievals tend to agree better at lower frequencies than at higher ones, with systematic differences ranging 1%-4% (3-12 K) over desert and 1%-7% (3-20 K) over rainforest. The random errors within each retrieval dataset are in the range of 0.5%-2% (2-6 K). In particular, at 85.5/89.0 GHz, there are very large differences between the different retrieval datasets, and within each retrieval dataset itself. Further investigation reveals that these differences are most likely caused by rain/cloud contamination, which can lead to random errors up to 10-17 K under the most severe conditions.

A Comparison of Microwave Window Channel Retrieved and Forward-Modeled Emissivities Over the U.S. Southern Great Plains

An accurate understanding of land surface emissivity in terms of associated surface properties is necessary for improved passive microwave remote sensing of the atmosphere, including water vapor, clouds, and precipitation, over land. In an effort to advance this understanding, emissivities are calculated for a 5 ° latitude by 5 ° longitude region in the U.S. Southern Great Plains using a combination of land surface model and physical emissivity model. Results are compared to retrieved values from the Advanced Microwave Scanning Radiometer-Earth Observing System passive microwave observations for cloud-free scenes over a six-year period. The resulting emissivities are compared in the context of surface properties including surface temperature, leaf area index (LAI), soil moisture, and precipitation. The comparison confirms that lower frequency channels respond most directly to the surface soil and its dielectric properties. Differences between retrieved and modeled emissivities are generally lower than 2%-3% and appear to be a function of soil moisture and LAI at frequencies less than 37 GHz. Agreement is better for the vertical polarization channels. At 89 GHz, a large difference is present between retrieved and modeled emissivities in both mean and magnitude of variability, particularly in the summer months. Problems are likely present at higher microwave frequencies in both the retrieved and modeled products, including the inability of the emissivity model to represent liquid water in the form of dew or precipitation interception on the vegetation canopy.

Assessing the Impact of L-Band Observations on Drought and Flood Risk Estimation: A Decision-Theoretic Approach in an OSSE Environment

AbstractObserving system simulation experiments (OSSEs) are often conducted to evaluate the worth of existing data and data yet to be collected from proposed new missions. As missions increasingly require a broader “Earth systems” focus, it is important that the OSSEs capture the potential benefits of the observations on end-use applications. Toward this end, the results from the OSSEs must also be evaluated with a suite of metrics that capture the value, uncertainty, and information content of the observations while factoring in both science and societal impacts. This article presents a soil moisture OSSE that employs simulated L-band measurements and assesses its utility toward improving drought and flood risk estimates using the NASA Land Information System (LIS). A decision-theory-based analysis is conducted to assess the economic utility of the observations toward improving these applications. The results suggest that the improvements in surface soil moisture, root-zone soil moisture, and total runof...

Performance Metrics, Error Modeling, and Uncertainty Quantification

AbstractA common set of statistical metrics has been used to summarize the performance of models or measurements—the most widely used ones being bias, mean square error, and linear correlation coefficient. They assume linear, additive, Gaussian errors, and they are interdependent, incomplete, and incapable of directly quantifying uncertainty. The authors demonstrate that these metrics can be directly derived from the parameters of the simple linear error model. Since a correct error model captures the full error information, it is argued that the specification of a parametric error model should be an alternative to the metrics-based approach. The error-modeling methodology is applicable to both linear and nonlinear errors, while the metrics are only meaningful for linear errors. In addition, the error model expresses the error structure more naturally, and directly quantifies uncertainty. This argument is further explained by highlighting the intrinsic connections between the performance metrics, the erro...

Bayesian approach to contaminant source characterization in water distribution systems: adaptive sampling framework

Bayesian analysis can yield a probabilistic contaminant source characterization conditioned on available sensor data and accounting for system stochastic processes. This paper is based on a previously proposed Markov chain Monte Carlo (MCMC) approach tailored for water distribution systems and incorporating stochastic water demands. The observations can include those from fixed sensors and, the focus of this paper, mobile sensors. Decision makers, such as utility managers, need not wait until new observations are available from an existing sparse network of fixed sensors. This paper addresses a key research question: where is the best location in the network to gather additional measurements so as to maximize the reduction in the source uncertainty? Although this has been done in groundwater management, it has not been well addressed in water distribution networks. In this study, an adaptive framework is proposed to guide the strategic placement of mobile sensors to complement the fixed sensor network. MCMC is the core component of the proposed adaptive framework, while several other pieces are indispensable: Bayesian preposterior analysis, value of information criterion and the search strategy for identifying an optimal location. Such a framework is demonstrated with an illustrative example, where four candidate sampling locations in the small water distribution network are investigated. Use of different value-of-information criteria reveals that while each may lead to different outcomes, they share some common characteristics. The results demonstrate the potential of Bayesian analysis and the MCMC method for contaminant event management.

An examination of methods for estimating land surface microwave emissivity

Land surface emissivity is a critical variable for the passive microwave-based remote sensing of the land and atmosphere. Driven by the Global Precipitation Measurement mission, we implemented and evaluated a variety of approaches for quantitative estimation of land surface emissivity and its variability, within a well-defined common framework. These approaches fall into three classes: physical modeling, statistical modeling, and a hybrid of physical and statistical modeling. Every approach is subject to evaluation against retrieved emissivity over a large area in the Southern Great Plains for a period of 2 years. Physical modeling, based on two radiative transfer models coupled to a land surface modeling framework, produced reasonable estimates, with channel- and polarization-dependent errors. Calibration of these models with historical data substantially improved their performance at lower frequencies. The statistical method was tested with five different regression models, and each of them consistently outperformed physical models by about 50%. The best statistical model had an average error of 0.9–2.1%. These statistical models were slightly improved when empirical orthogonal function analysis was incorporated in the emissivity data. The hybrid approach produced errors between the uncalibrated and calibrated physical model errors. In addition to their predictive performance, other aspects of each approachs strengths and weaknesses are discussed.

Calibration to Improve Forward Model Simulation of Microwave Emissivity at GPM Frequencies Over the U.S. Southern Great Plains

Better estimation of land surface microwave emissivity (MWE) promises to improve overland precipitation retrievals in the Global Precipitation Measurement era. Forward models of land MWE are available but have suffered from poor parameter specification and limited testing. Here, forward models are calibrated, and the accompanying change in predictive power is evaluated. With inputs (e.g., soil moisture) from the Noah land surface model and applying Moderate Resolution Imaging Spectroradiometer leaf area index data, two microwave emissivity models (MEMs) are tested, namely, the Community Radiative Transfer Model and the Community Microwave Emission Model. The calibration is conducted with the National Aeronautics and Space Administration Land Information System parameter estimation subsystem using Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E)-based emissivity retrievals for the calibration data set. The extent of agreement between the modeled and retrieved estimates is evaluated using the AMSR-E retrievals for a separate seven-year validation period. Results indicate that calibration can significantly improve the agreement, simulating emissivity with an across-channel average root-mean-square difference (RMSD) of about 0.013 or about 20% lower than if relying on daily estimates based on climatology. The results also indicate that calibration of the MEM alone, as was done in prior studies, results in as much as 12% higher across-channel average RMSD, as compared with the joint calibration of the land surface and microwave emissivity models. It remains as future work to assess the extent to which the improvements in emissivity estimation translate into improvements in precipitation retrieval accuracy.