@article{fuentes_kittel_nychka_2006, title={Sensitivity of ecological models to their climate drivers: Statistical ensembles for forcing}, volume={16}, ISSN={["1939-5582"]}, DOI={10.1890/04-1157}, abstractNote={Global and regional numerical models for terrestrial ecosystem dynamics require fine spatial resolution and temporally complete historical climate fields as input variables. However, because climate observations are unevenly spaced and have incomplete records, such fields need to be estimated. In addition, uncertainty in these fields associated with their estimation are rarely assessed. Ecological models are usually driven with a geostatistical model's mean estimate (kriging) of these fields without accounting for this uncertainty, much less evaluating such errors in terms of their propagation in ecological simulations. We introduce a Bayesian statistical framework to model climate observations to create spatially uniform and temporally complete fields, taking into account correlation in time and space, spatial heterogeneity, lack of normality, and uncertainty about all these factors. A key benefit of the Bayesian model is that it generates uncertainty measures for the generated fields. To demonstrate this method, we reconstruct historical monthly precipitation fields (a driver for ecological models) on a fine resolution grid for a climatically heterogeneous region in the western United States. The main goal of this work is to evaluate the sensitivity of ecological models to the uncertainty associated with prediction of their climate drivers. To assess their numerical sensitivity to predicted input variables, we generate a set of ecological model simulations run using an ensemble of different versions of the reconstructed fields. We construct such an ensemble by sampling from the posterior predictive distribution of the climate field. We demonstrate that the estimated prediction error of the climate field can be very high. We evaluate the importance of such errors in ecological model experiments using an ensemble of historical precipitation time series in simulations of grassland biogeochemical dynamics with an ecological numerical model, Century. We show how uncertainty in predicted precipitation fields is propagated into ecological model results and that this propagation had different modes. Depending on output variable, the response of model dynamics to uncertainty in inputs ranged from uncertainty in outputs that matched that of inputs to those that were muted or that were biased, as well as uncertainty that was persistent in time after input errors dropped.}, number={1}, journal={ECOLOGICAL APPLICATIONS}, author={Fuentes, M and Kittel, TGF and Nychka, D}, year={2006}, month={Feb}, pages={99–116} }
 @article{davis_nychka_bailey_2000, title={A comparison of regional oxidant model (ROM) output with observed ozone data}, volume={34}, ISSN={["1352-2310"]}, DOI={10.1016/S1352-2310(99)00424-0}, abstractNote={The output from the regional oxidant model (ROM) is compared to observed ozone over northern Illinois for June, July and August 1987. The 8-h daily average ozone at the ozone monitoring stations is interpolated to the ROM grid cells using a spatial statistical method. Differences between the model output and spatial predictions are compared at three levels of spatial averaging (with approximate scales of 19, 100, 400 km) and three levels of temporal averaging (daily, weekly, 3 months). In addition two ozone monitoring stations are paired with weather stations and with ROM cells in order to investigate the performance of ROM as a function of meteorological conditions. For daily values the root mean squared error (RMSE) between the ROM values and those predicted from the monitoring network varies between 14 and 25 ppb with the largest discrepancies occurring near Lake Michigan. Weekly averages reduce the RMSE by approximately 30% but spatial aggregation is not helpful in improving the agreement. The difference between ROM ozone predictions and the observed ozone at two paired sites depends most strongly on temperature and to a lesser extent on dew point temperature. The R2 from linear regressions is approximately 35%. An examination of the synoptic-scale and meso-scale weather patterns during this period indicates that ROM is sensitive to dynamic situations such as a frontal passage.}, number={15}, journal={ATMOSPHERIC ENVIRONMENT}, author={Davis, JM and Nychka, D and Bailey, B}, year={2000}, pages={2413–2423} }
 @article{hu_hall_nychka_2000, title={A nonparametric approach to stochastic discount factor estimation}, volume={14}, number={2000}, journal={Advances in Econometrics}, author={Hu, F. and Hall, A. R. and Nychka, D.}, year={2000}, pages={155–176} }
 @article{huang_nychka_2000, title={A nonparametric multiple choice method within the random utility framework}, volume={97}, ISSN={["0304-4076"]}, DOI={10.1016/S0304-4076(99)00072-X}, abstractNote={Many researchers use categorical data analysis to recover individual consumption preferences, but the standard discrete choice models require restrictive assumptions. To improve the flexibility of discrete choice data analysis, we propose a nonparametric multiple choice model that applies the penalized likelihood method within the random utility framework. We show that the deterministic component of the random utility function in the model is a cubic smoothing spline function. The method subsumes the conventional conditional logit model (McFadden, 1973, in: Zarembka, P., (Ed.), Frontiers in Econometrics) as a special case. In this paper, we present the model, describe the estimator, provide the computational algorithm of the model, and demonstrate the model by applying it to nonmarket valuation of recreation sites.}, number={2}, journal={JOURNAL OF ECONOMETRICS}, author={Huang, JC and Nychka, DW}, year={2000}, month={Aug}, pages={207–225} }
 @article{tsai_brownie_nychka_pollock_1999, title={Smoothing hazard functions for telemetry survival data in wildlife studies}, volume={46}, number={1999}, journal={Bird Study}, author={Tsai, K. and Brownie, C. and Nychka, D. W. and Pollock, K. H.}, year={1999}, pages={47–54} }
 @article{royle_nychka_1998, title={An algorithm for the construction of spatial coverage designs with implementation in SPLUS}, volume={24}, ISSN={["0098-3004"]}, DOI={10.1016/s0098-3004(98)00020-x}, abstractNote={Space-filling “coverage” designs are spatial sampling plans which optimize a distance-based criterion. Because they do not depend on the covariance structure of the process to be sampled, coverage designs are computed more efficiently than designs that are optimal for mean-squared-error criteria. This paper presents an efficient algorithm for the construction of coverage designs and evaluates its performance in terms of computation time and effectiveness at finding “good” designs. Results suggest that near-optimal designs for reasonably large problems can be computed efficiently. The algorithm is implemented in the statistical programming language SPLUS and examples of the construction of coverage designs are given involving an existing network of ozone monitoring sites.}, number={5}, journal={COMPUTERS & GEOSCIENCES}, author={Royle, JA and Nychka, D}, year={1998}, month={Jun}, pages={479–488} }
 @article{davis_eder_nychka_yang_1998, title={Modeling the effects of meteorology on ozone in Houston using cluster analysis and generalized additive models}, volume={32}, ISSN={["1352-2310"]}, DOI={10.1016/S1352-2310(98)00008-9}, abstractNote={This paper compares the results from a single-stage clustering technique (average linkage) with those of a two-stage technique (average linkage then k-means) as part of an objective meteorological classification scheme designed to better elucidate ozone’s dependence on meteorology in the Houston, Texas, area. When applied to twelve years of meteorological data (1981–1992), each clustering technique identified seven statistically distinct meteorological regimes. The majority of these regimes exhibited significantly different daily 1 h maximum ozone (O3) concentrations, with the two-stage approach resulting in a better segregation of the mean concentrations when compared to the single-stage approach. Both approaches indicated that the largest daily 1 h maximum concentrations are associated with migrating anticyclones that occur most often during spring and summer, and not with the quasi-permanent Bermuda High that often dominates the southeastern United States during the summer. As a result, maximum ozone concentrations are just as likely during the months of April, May, September and October as they are during the summer months. Generalized additive models were then developed within each meteorological regime in order to identify those meteorological covariates most closely associated with O3 concentrations. Three surface wind covariates: speed, and the u and v components were selected nearly unanimously in those meteorological regimes dominated by anticyclones, indicating the importance of transport within these O3 conducive meteorological regimes.}, number={14-15}, journal={ATMOSPHERIC ENVIRONMENT}, author={Davis, JM and Eder, BK and Nychka, D and Yang, Q}, year={1998}, month={Aug}, pages={2505–2520} }
 @article{ellner_bailey_bobashev_gallant_grenfell_nychka_1998, title={Noise and nonlinearity in measles epidemics: Combining mechanistic and statistical approaches to population modeling}, volume={151}, ISSN={["1537-5323"]}, DOI={10.1086/286130}, abstractNote={We present and evaluate an approach to analyzing population dynamics data using semimechanistic models. These models incorporate reliable information on population structure and underlying dynamic mechanisms but use nonparametric surface‐fitting methods to avoid unsupported assumptions about the precise form of rate equations. Using historical data on measles epidemics as a case study, we show how this approach can lead to better forecasts, better characterizations of the dynamics, and a better understanding of the factors causing complex population dynamics relative to either mechanistic models or purely descriptive statistical time‐series models. The semimechanistic models are found to have better forecasting accuracy than either of the model types used in previous analyses when tested on data not used to fit the models. The dynamics are characterized as being both nonlinear and noisy, and the global dynamics are clustered very tightly near the border of stability (dominant Lyapunov exponent λ ≈ 0). However, locally in state space the dynamics oscillate between strong short‐term stability and strong short‐term chaos (i.e., between negative and positive local Lyapunov exponents). There is statistically significant evidence for short‐term chaos in all data sets examined. Thus the nonlinearity in these systems is characterized by the variance over state space in local measures of chaos versus stability rather than a single summary measure of the overall dynamics as either chaotic or nonchaotic.}, number={5}, journal={AMERICAN NATURALIST}, author={Ellner, SP and Bailey, BA and Bobashev, GV and Gallant, AR and Grenfell, BT and Nychka, DW}, year={1998}, month={May}, pages={425–440} }
 @inproceedings{eder_davis_nychka_1997, title={The impact of meteorology on ozone in Huston}, booktitle={EPA/A&WMA International Symposium on Measurement of Toxic and Related Air Pollutants (1992 May 4-9: Durham, N.C.) Measurement of toxic and related air pollutants: Proceedings of the 1992 EPA/A&WMA International Symposium}, publisher={Pittsburgh, Pa.: A&WMA}, author={Eder, B. K. and Davis, J. M. and Nychka, D.}, year={1997}, pages={204–214} }