Local Covariance Research Articles

A Nonvariational Consistent Hybrid Ensemble Filter

Abstract A consistent hybrid ensemble filter (CHEF) for using hybrid forecast error covariance matrices that linearly combine aspects of both climatological and flow-dependent matrices within a nonvariational ensemble data assimilation scheme is described. The CHEF accommodates the ensemble data assimilation enhancements of (i) model space ensemble covariance localization for satellite data assimilation and (ii) Hodyss’s method for improving accuracy using ensemble skewness. Like the local ensemble transform Kalman filter (LETKF), the CHEF is computationally scalable because it updates local patches of the atmosphere independently of others. Like the sequential ensemble Kalman filter (EnKF), it serially assimilates batches of observations and uses perturbed observations to create ensembles of analyses. It differs from the deterministic (no perturbed observations) ensemble square root filter (ESRF) and the EnKF in that (i) its analysis correction is unaffected by the order in which observations are assimilated even when localization is required, (ii) it uses accurate high-rank solutions for the posterior error covariance matrix to serially assimilate observations, and (iii) it accommodates high-rank hybrid error covariance models. Experiments were performed to assess the effect on CHEF and ESRF analysis accuracy of these differences. In the case where both the CHEF and the ESRF used tuned localized ensemble covariances for the forecast error covariance model, the CHEF’s advantage over the ESRF increased with observational density. In the case where the CHEF used a hybrid error covariance model but the ESRF did not, the CHEF had a substantial advantage for all observational densities.

Local spectral variability features for speaker verification

Speaker verification techniques neglect the short-time variation in the feature space even though it contains speaker related attributes. We propose a simple method to capture and characterize this spectral variation through the eigenstructure of the sample covariance matrix. This covariance is computed using sliding window over spectral features. The newly formulated feature vectors representing local spectral variations are used with classical and state-of-the-art speaker recognition systems. Results on multiple speaker recognition evaluation corpora reveal that eigenvectors weighted with their normalized singular values are useful in representing local covariance information. We have also shown that local variability features can be extracted using mel frequency cepstral coefficients (MFCCs) as well as using three recently developed features: frequency domain linear prediction (FDLP), mean Hilbert envelope coefficients (MHECs) and power-normalized cepstral coefficients (PNCCs). Since information conveyed in the proposed feature is complementary to the standard short-term features, we apply different fusion techniques. We observe considerable relative improvements in speaker verification accuracy in combined mode on text-independent (NIST SRE) and text-dependent (RSR2015) speech corpora. We have obtained up to 12.28% relative improvement in speaker recognition accuracy on text-independent corpora. Conversely in experiments on text-dependent corpora, we have achieved up to 40% relative reduction in EER. To sum up, combining local covariance information with the traditional cepstral features holds promise as an additional speaker cue in both text-independent and text-dependent recognition.

EVALUATION OF SPATIAL DEPENDENCE OF CORN (Zea mays L.) MORPHOMETRIC CHARACTERISTICS ON EDAPHIC PROPERTIES

Spatial patterns of maize morphometrics traits interrelation variability with geographically weighted principal components analysis at large-scale level have been revealed. Spatial variability of covariation structures which describes interrelation between morphometrics indicators and density of standing of maize have been established. The global pattern of interrelation obtained by means of the classical principal component analysis has been shown as not to be identical to local covariation structures. Local covariation structures which found with geographically weighted principal components analysis within an optimum kernel bandwidth are characterized by much higher level of nonrandom variability which is described by first three principal components. The part of a dispersion which specifies in a coordination of morphological structures is characterized by natural spatial trends of a variation. Local covariation structures form spatially natural patterns of the placing. Feature of these structures is quantitative redistribution of those values or other signs within the limits of enough invariant configurations. The continuity covariation structures in a qualitative sense at various scale levels (global and local) but with local quantitative specificity is shown. This specificity is shown in prevalence of this or that indicator as basic marker main components at local level. Revealing of spatial patterns covariation structures puts a problem of understanding of the nature of this spatial regularity. Keywords: principal components analysis, morphometric traits, spatial variability

CONSTRUCTING A FLEXIBLE LIKELIHOOD FUNCTION FOR SPECTROSCOPIC INFERENCE

We present a modular, extensible likelihood framework for spectroscopic inference based on synthetic model spectra. The subtraction of an imperfect model from a continuously sampled spectrum introduces covariance between adjacent datapoints (pixels) into the residual spectrum. For the high signal-to-noise data with large spectral range that is commonly employed in stellar astrophysics, that covariant structure can lead to dramatically underestimated parameter uncertainties (and, in some cases, biases). We construct a likelihood function that accounts for the structure of the covariance matrix, utilizing the machinery of Gaussian process kernels. This framework specifically address the common problem of mismatches in model spectral line strengths (with respect to data) due to intrinsic model imperfections (e.g., in the atomic/molecular databases or opacity prescriptions) by developing a novel local covariance kernel formalism that identifies and self-consistently downweights pathological spectral line "outliers." By fitting many spectra in a hierarchical manner, these local kernels provide a mechanism to learn about and build data-driven corrections to synthetic spectral libraries. An open-source software implementation of this approach is available at http://iancze.github.io/Starfish, including a sophisticated probabilistic scheme for spectral interpolation when using model libraries that are sparsely sampled in the stellar parameters. We demonstrate some salient features of the framework by fitting the high resolution $V$-band spectrum of WASP-14, an F5 dwarf with a transiting exoplanet, and the moderate resolution $K$-band spectrum of Gliese 51, an M5 field dwarf.

Effect of spatial sampling from European flux towers for estimating carbon and water fluxes with artificial neural networks

AbstractEmpirical modeling approaches are frequently used to upscale local eddy covariance observations of carbon, water, and energy fluxes to regional and global scales. The predictive capacity of such models largely depends on the data used for parameterization and identification of input‐output relationships, while prediction for conditions outside the training domain is generally uncertain. In this work, artificial neural networks (ANNs) were used for the prediction of gross primary production (GPP) and latent heat flux (LE) on local and European scales with the aim to assess the portion of uncertainties in extrapolation due to sample selection. ANNs were found to be a useful tool for GPP and LE prediction, in particular for extrapolation in time (mean absolute error MAE for GPP between 0.53 and 1.56 gC m−2 d−1). Extrapolation in space in similar climatic and vegetation conditions also gave good results (GPP MAE 0.7–1.41 gC m−2 d−1), while extrapolation in areas with different seasonal cycles and controlling factors (e.g., the tropical regions) showed noticeably higher errors (GPP MAE 0.8–2.09 gC m−2 d−1). The distribution and the number of sites used for ANN training had a remarkable effect on prediction uncertainty in both, regional GPP and LE budgets and their interannual variability. Results obtained show that for ANN upscaling for continents with relatively small networks of sites, the error due to the sampling can be large and needs to be considered and quantified. The analysis of the spatial variability of the uncertainty helped to identify the meteorological drivers driving the uncertainty.

Improving high resolution emission inventories with local proxies and urban eddy covariance flux measurements

Emission inventories are the fundamental official data on atmospheric emissions of pollutants and greenhouse gases at a variety of spatial and temporal scales worldwide. This study makes use of direct CO2 emission measurements made with the eddy covariance technique over a completely urbanized area, with no confounding effect of vegetation, where emissions are mostly controlled by natural gas combustion processes and road traffic. Objectives are: i) to validate top-down spatially and temporally disaggregated emission inventories at yearly, monthly, weekly and hourly time scales; ii) to quantify the improvement achieved in official inventories when replacing built-in temporal disaggregation proxies with customized proxies based on local data of road traffic and natural gas consumption. We demonstrate that the overall performance of official inventory at yearly scale is rather good with an emission of 3.08 g CO2 m−2 h−1 against a measured emission of 3.21 ± 0.12 g CO2 m−2 h−1. When temporally disaggregating annual emissions, the agreement between inventory and observations always significantly improves when using local proxies, by 47% (from 0.70 to 0.37 g CO2 m−2 h−1 RMSE) at monthly scale, by 26% (from 0.58 to 0.43 g CO2 m−2 h−1 RMSE) at weekly scale, and by 32% (from 1.26 to 0.85 g CO2 m−2 h−1 RMSE), at hourly scale. The validity of this analysis goes beyond CO2 since the temporal proxies used by the inventories mimic the intensity of specific emission processes, therefore species emitted in the same processes as CO2, would benefit from the improved parameterization of temporal proxies shown here. These results indicate that effort should be put into developing improved temporal proxies based on local rather than national scale data, that can better mimic site dependent behaviors.

Dynamical Locality of the Free Maxwell Field

We consider the non-interacting source-free Maxwell field, described both in terms of the vector potential and the field strength. Starting from the classical field theory on contractible globally hyperbolic spacetimes, we extend the classical field theory to general globally hyperbolic spacetimes in two ways to obtain a "universal" theory and a "reduced" theory. The quantum field theory in terms of the unital $*$-algebra of the smeared quantum field is then obtained by an application of a suitable quantisation functor. We show that the universal theories fail local covariance and dynamical locality owing to the possibility of having non-trivial radicals in the classical and non-trivial centres in the quantum case. The reduced theories are both locally covariant and dynamically local. These models provide new examples relevant to the discussion of how theories should be formulated so as to describe the same physics in all spacetimes.

Decision fusion for dual-window-based hyperspectral anomaly detector

In hyperspectral anomaly detection, the dual-window-based detector is a widely used technique that employs two windows to capture nonstationary statistics of anomalies and back- ground. However, its detection performance is usually sensitive to the choice of window sizes and suffers from inappropriate window settings. In this work, a decision-fusion approach is pro- posed to alleviate such sensitivity by merging the results from multiple detectors with different window sizes. The proposed approach is compared with the classic Reed-Xiaoli (RX) algorithm as well as kernel RX (KRX) using two real hyperspectral data. Experimental results demonstrate that it outperforms the existing detectors, such as RX, KRX, and multiple-window-based RX. The overall detection framework is suitable for parallel computing, which can greatly reduce computational time when processing large-scale remote sensing image data. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or repro- duction of this work in whole or in part requires full attribution of the original publication, including its DOI. (DOI: 10.1117/1.JRS.9.097297) tional probability density functions under the two hypotheses (without and with anomaly) are assumed to be Gaussian. The solution turns out to be an adaptive Mahalanobis distance between the pixel under test and the local background. It is preferred to use local background to capture nonstationary statistics, and its advantage of using a global background covariance matrix has been demonstrated in the literature. 11-13 The RX detector has become the benchmark of anomaly detection algorithms in HSI. Obviously, the key to success is an appropriate estimate of a local background covariance matrix for effective background suppression. An adaptive RX detector employs a dual-window strategy: the inner window is slightly larger than the pixel size, the outer window is even larger than the inner one, and only the samples in the outer region (i.e., between the frames of inner and outer windows) are used to estimate the background covariance matrix to avoid the use of the potential anomalous pixels. Intuitively, the number of pixels in the outer region (related to the sizes of inner and outer windows) should be more than the number of bands so that the resulting covariance matrix can be full-rank for inverse matrix operation. However, even when the covariance matrix is ill-rank, its inversion can still be computed by several strategies, such

A comparison of strategies for estimation of ultrafine particle number concentrations in urban air pollution monitoring networks

We propose three estimation strategies (local, remote and mixed) for ultrafine particles (UFP) at three sites in an urban air pollution monitoring network. Estimates are obtained through Gaussian process regression based on concentrations of gaseous pollutants (NOx, O3, CO) and UFP. As local strategy, we use local measurements of gaseous pollutants (local covariates) to estimate UFP at the same site. As remote strategy, we use measurements of gaseous pollutants and UFP from two independent sites (remote covariates) to estimate UFP at a third site. As mixed strategy, we use local and remote covariates to estimate UFP. The results suggest: UFP can be estimated with good accuracy based on NOx measurements at the same location; it is possible to estimate UFP at one location based on measurements of NOx or UFP at two remote locations; the addition of remote UFP to local NOx, O3 or CO measurements improves models' performance.

Local environmental covariates are important for predicting fire history from tree stem diameters

In fire-prone landscapes, knowing when vegetation was last burnt is important for understanding how species respond to fire and to develop effective fire management strategies. However, fire history is often incomplete or non-existent. We developed a fire-age prediction model for two mallee woodland tree species in southern Australia. The models were based on stem diameters from ~1172 individuals surveyed along 87 transects. Time since fire accounted for the greatest proportion of the explained variation in stem diameter for our two mallee tree species but variation in mean stem diameters was also influenced by local environmental factors. We illustrate a simple tool that enables time since fire to be predicted based on stem diameter and local covariates. We tested our model against new data but it performed poorly with respect to the mapped fire history. A combination of different covariate effects, variation in among-tree competition, including above- and below-ground competition, and unreliable fire history may have contributed to poor model performance. Understanding how the influence of covariates on stem diameter growth varies spatially is critical for determining the generality of models that predict time since fire. Models that were developed in one region may need to be independently verified before they can be reliably applied in new regions.

Comparison of Hybrid-4DEnVar and Hybrid-4DVar Data Assimilation Methods for Global NWP

Abstract The Met Office has developed an ensemble-variational data assimilation method (hybrid-4DEnVar) as a potential replacement for the hybrid four-dimensional variational data assimilation (hybrid-4DVar), which is the current operational method for global NWP. Both are four-dimensional variational methods, using a hybrid combination of a fixed climatological model of background error covariances with localized covariances from an ensemble of current forecasts designed to describe the structure of “errors of the day.” The fundamental difference between the methods is their modeling of the time evolution of errors within each data assimilation window: 4DVar uses a linear model and its adjoint and 4DEnVar uses a localized linear combination of nonlinear forecasts. Both hybrid-4DVar and hybrid-4DEnVar beat their three-dimensional versions, which are equivalent, in NWP trials. With settings based on the current operational system, hybrid-4DVar performs better than hybrid-4DEnVar. Idealized experiments designed to compare the time evolution of covariances in the methods are described: the basic 4DEnVar represents the evolution of ensemble errors as well as 4DVar. However, 4DVar also represents the evolution of errors from the climatological covariances, whereas 4DEnVar does not. This difference is the main cause of the superiority of hybrid-4DVar. Another difference is that the authors’ 4DVar explicitly penalizes rapid variations in the analysis increment trajectory, while the authors’ 4DEnVar contains no dynamical constaints on imbalance. The authors describe a four-dimensional incremental analysis update (4DIAU) method that filters out the high-frequency oscillations introduced by the poorly balanced 4DEnVar increments. Possible methods for improving hybrid-4DEnVar are discussed.

Sampling from Determinantal Point Processes for Scalable Manifold Learning.

High computational costs of manifold learning prohibit its application for large datasets. A common strategy to overcome this problem is to perform dimensionality reduction on selected landmarks and to successively embed the entire dataset with the Nyström method. The two main challenges that arise are: (i) the landmarks selected in non-Euclidean geometries must result in a low reconstruction error, (ii) the graph constructed from sparsely sampled landmarks must approximate the manifold well. We propose to sample the landmarks from determinantal distributions on non-Euclidean spaces. Since current determinantal sampling algorithms have the same complexity as those for manifold learning, we present an efficient approximation with linear complexity. Further, we recover the local geometry after the sparsification by assigning each landmark a local covariance matrix, estimated from the original point set. The resulting neighborhood selection .based on the Bhattacharyya distance improves the embedding of sparsely sampled manifolds. Our experiments show a significant performance improvement compared to state-of-the-art landmark selection techniques on synthetic and medical data.

A Probabilistic Approach to Adaptive Covariance Localization for Serial Ensemble Square Root Filters

Abstract This study proposes a variational approach to adaptively determine the optimum radius of influence for ensemble covariance localization when uncorrelated observations are assimilated sequentially. The covariance localization is commonly used by various ensemble Kalman filters to limit the impact of covariance sampling errors when the ensemble size is small relative to the dimension of the state. The probabilistic approach is based on the premise of finding an optimum localization radius that minimizes the distance between the Kalman update using the localized sampling covariance versus using the true covariance, when the sequential ensemble Kalman square root filter method is used. The authors first examine the effectiveness of the proposed method for the cases when the true covariance is known or can be approximated by a sufficiently large ensemble size. Not surprisingly, it is found that the smaller the true covariance distance or the smaller the ensemble, the smaller the localization radius that is needed. The authors further generalize the method to the more usual scenario that the true covariance is unknown but can be represented or estimated probabilistically based on the ensemble sampling covariance. The mathematical formula for this probabilistic and adaptive approach with the use of the Jeffreys prior is derived. Promising results and limitations of this new method are discussed through experiments using the Lorenz-96 system.

Blind Source Separation for Spatial Compositional Data

In regional geochemistry rock, sediment, soil, plant or water samples, collected in a certain region, are analyzed for concentrations of chemical elements. The observations are thus usually high dimensional, spatially dependent and of compositional nature. In this paper, a novel blind source separation approach for spatially dependent data is suggested. For the analysis, it is assumed that the multivariate observations are linear combinations or mixtures of latent components and that the spatial processes for these latent components are second order stationary and uncorrelated. In the present approach, the latent components are then recovered by simultaneously diagonalizing the covariance matrix and a local covariance (correlation) matrix. This method can be easily applied also in the context of compositional data after appropriate data transformations. The components obtained in this way are uncorrelated and easily interpretable, and can be used for dimension reduction and for visual presentation of different features of the data. To demonstrate the usefulness of the new method, the KOLA data are reanalyzed using the new procedure and the results are compared to the results coming from marginal principal component analysis and independent component analysis that ignore spatial dependence.

Kernel-Based MMSE Multimedia Signal Reconstruction and Its Application to Spatial Error Concealment

This paper proposes a novel approach for multimedia signal reconstruction based on kernel density estimation (KDE). We make use of a vector formalism in which vectors consist of a first subvector containing a set of missing samples and a second one containing a set of available context samples. The missing subvector is reconstructed by a minimum mean square error estimator which employs a probability density function (pdf) obtained by KDE. As in any kernel-based method, the main issue to deal with is the estimation of an appropriate kernel bandwidth. We propose an adaptive procedure for bandwidth estimation (BE) especially conceived for signal reconstruction. Thus, unlike general KDE or kernel-based regression, which try to obtain a general fit, the focus of this BE procedure is on the specific missing subvector. Also, in order to exploit local signal correlations, our BE proposal adopts a scaling approach in which the bandwidth is computed as the local covariance matrix scaled by two factors. These two scale factors are obtained by minimization of two different approximations to the reconstruction error. The resulting reconstruction methodology is tested on a spatial error concealment (EC) application in which intracoded images have been transmitted through an error prone channel. The experimental results show the superiority of the proposed approach over a wide range of existing EC techniques.

Abstract 4263: Identifying somatic mutation hotspots across protein family alignments

Abstract In cancer, somatic driver mutations often target “hotspots” of paralogous residues across evolutionarily related members of a single protein family. These hotspots usually confer a dominant oncogenic function, such as constitutive catalytic activity or binding. These hotspots generally represent the most “druggable” genetic alterations. Examples of protein families targeted in such a manner include the Ras and receptor Tyrosine Kinase family. Protein family hotspots are not directly assessed by standard computational approaches for statistically nominating genes under positive somatic selection, such as MutSig, InVex, or MuSic. These methods only assess single genes for statistical enrichment of total mutation burden or the presence of mutation hotspots. As a result, such approaches are only powered to detect protein family hotspots in cases where each (or at least one) protein family member is very frequently mutated (e.g. KRAS, NRAS, HRAS). They are not powered to detect hotspots in highly functionally redundant families, of which each member may be only mutated in <<1% of a tumor type. To directly identify such protein-family hotspots, we have developed a computational framework, MutPfam. This framework superimposes mutation data onto Pfam (http://pfam.sanger.ac.uk/) (protein family and subfamily multiple sequence alignments and identifies statistically significant hotspots. We demonstate our method on protein families involved in the catalysis of small-molecular catabolism using the TCGA Pan-Cancer dataset (https://tcga-data.nci.nih.gov/) comprising somatic mutation calls for 4608 whole-exome sequenced patients and 21 tumor types. As we demonstrate, our framework enables the de novo discovery of mutation hotspots across protein families. Though developed for somatic mutation analysis, this framework could also be applied to the analysis of inherited mutations. Computational issues meriting future investigation include (1) alternate definitions of protein alignments and sub-families (e.g. involving three dimensional structure, biophysical constraints) (2) improved statistical modeling of the local background mutation rate (incorporating local genomic covariates that mediate heterogeneity) (3) incorporation of variant functional impact scores for missense variants. Citation Format: Marcin Imielinski, Charles Du, Matthew Meyerson. Identifying somatic mutation hotspots across protein family alignments. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 4263. doi:10.1158/1538-7445.AM2014-4263

System-wide tail comovements: A bootstrap test for cojump identification on the S&P 500, US bonds and currencies

This paper studies bivariate tail comovements on financial markets that are of crucial importance for the world economy: the S&P 500, US bonds, and currencies. We propose to study that form of dependence under the lens of cojump identification in a bivariate Brownian semimartingale with idiosyncratic jumps, as well as cojumps. Whereas univariate jump identification has been widely studied in the high-frequency data literature, the multivariate literature on cojump identification is more recent and scarcer. Cojump identification is of interest, as it may identify comovements which are not trivially visible in a univariate setting. That is, price changes can be small relative to local variation, but still abnormal relative to local covariation. This paper investigates how simple parametric bootstrapping of the product of assets' intraday returns can help detect cojumps in a multivariate Brownian semi-martingale with both idiosyncratic jumps and cojumps. In particular, we investigate how to disentangle idiosyncratic jumps from common jumps at an intraday level for pairs of assets. The approach is flexible, trivial to implement, and yields good power properties. It allows to shed new light on extreme dependence at the world economy level. We detect cojumps of heterogeneous size which are partly undetected with a univariate approach. We find an increased cojump intensity after the crisis on the S&P 500-US bonds pair before a return to normal.

A C*-Algebra for Quantized Principal U(1)-Connections on Globally Hyperbolic Lorentzian Manifolds

The aim of this work is to complete our program on the quantization of connections on arbitrary principal U(1)-bundles over globally hyperbolic Lorentzian manifolds. In particular, we show that one can assign via a covariant functor to any such bundle an algebra of observables which separates gauge equivalence classes of connections. The C*-algebra we construct generalizes the usual CCR-algebras, since, contrary to the standard field-theoretic models, it is based on a presymplectic Abelian group instead of a symplectic vector space. We prove a no-go theorem according to which neither this functor, nor any of its quotients, satisfies the strict axioms of general local covariance. As a byproduct, we prove that a morphism violates the locality axiom if and only if a certain induced morphism of cohomology groups is non-injective. We show then that, fixing any principal U(1)-bundle, there exists a suitable category of subbundles for which a quotient of our functor yields a quantum field theory in the sense of Haag and Kastler. We shall provide a physical interpretation of this feature and we obtain some new insights concerning electric charges in locally covariant quantum field theory.

Adaptive Bayesian Nonstationary Modeling for Large Spatial Datasets Using Covariance Approximations

Gaussian process models have been widely used in spatial statistics but face tremendous modeling and computational challenges for very large nonstationary spatial datasets. To address these challenges, we develop a Bayesian modeling approach using a nonstationary covariance function constructed based on adaptively selected partitions. The partitioned nonstationary class allows one to knit together local covariance parameters into a valid global nonstationary covariance for prediction, where the local covariance parameters are allowed to be estimated within each partition to reduce computational cost. To further facilitate the computations in local covariance estimation and global prediction, we use the full-scale covariance approximation (FSA) approach for the Bayesian inference of our model. One of our contributions is to model the partitions stochastically by embedding a modified treed partitioning process into the hierarchical models that leads to automated partitioning and substantial computational benefits. We illustrate the utility of our method with simulation studies and the global Total Ozone Matrix Spectrometer (TOMS) data. Supplementary materials for this article are available online.

Avian community composition associated with interactions between local and landscape habitat attributes

As human demand for ecosystem products increases, managers of landscapes used for commodity production require information about effects of management regimes on biological diversity. Landscape attributes, however, may moderate ecological responses to local-scale conservation and management actions. As a result, uniform application of local management prescriptions may yield variable biodiversity responses. We examined how interactions between local habitat structure and landscape forest cover were associated with avian community composition in the Ouachita Mountains, Arkansas, USA, 1995–1998. We used Bayesian hierarchical models to estimate occupancy for 63 breeding bird species, while accounting for variable detection with data collected from 1941 temporally replicated point count stations. Specifically, we estimated how interactions of four local habitat covariates (canopy cover of mature coniferous and hardwood trees, number of snags, and shrub cover) with percentage of mature hardwood forest at the landscape scale were associated with species occupancy and richness. Average predictive comparisons indicated that snag count and shrub cover had the strongest associations with species richness. Estimated associations for each of the four local forest cover variables was similar across all levels of landscape forest cover, suggesting weak or negligible interactions between these local measures and the landscape covariate. We found little support for our main prediction that local/landscape habitat interactions would be strongest at low levels of landscape forest cover (1–20%). Consequently, we suggest that forest managers consider prescriptions that result in a broad spatial distribution of heterogeneous habitat structural conditions (e.g., variation in understory cover and composition), irrespective of landscape context, to maintain a diverse avian breeding assemblage on landscapes in this region.