SEARCH · Engineering Papers

Results for “statistical”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 181 records · Page 10

All sources NASA NTRS DOE OSTI

NASA NTRS2001

Statistical Ensemble of Large Eddy Simulations

A statistical ensemble of large eddy simulations (LES) is run simultaneously for the same flow. The information provided by the different large scale velocity fields is used to propose an ensemble averaged version of the dynamic model. This produces local model parameters that only depend on the statistical properties of the flow. An important property of the ensemble averaged dynamic procedure is that it does not require any spatial averaging and can thus be used in fully inhomogeneous flows. Also, the ensemble of LES's provides statistics of the large scale velocity that can be used for building new models for the subgrid-scale stress tensor. The ensemble averaged dynamic procedure has been implemented with various models for three flows: decaying isotropic turbulence, forced isotropic turbulence, and the time developing plane wake. It is found that the results are almost independent of the number of LES's in the statistical ensemble provided that the ensemble contains at least 16 realizations.

Carati, Daniele↗

NASA NTRS2001

Robust Combining of Disparate Classifiers Through Order Statistics

Integrating the outputs of multiple classifiers via combiners or meta-learners has led to substantial improvements in several difficult pattern recognition problems. In this article we investigate a family of combiners based on order statistics, for robust handling of situations where there are large discrepancies in performance of individual classifiers. Based on a mathematical modeling of how the decision boundaries are affected by order statistic combiners, we derive expressions for the reductions in error expected when simple output combination methods based on the the median, the maximum and in general, the ith order statistic, are used. Furthermore, we analyze the trim and spread combiners, both based on linear combinations of the ordered classifier outputs, and show that in the presence of uneven classifier performance, they often provide substantial gains over both linear and simple order statistics combiners. Experimental results on both real world data and standard public domain data sets corroborate these findings.

Tumer, Kagan↗

NASA NTRS2002

Statistical Sampling of Tide Heights Study

The goal of the study was to determine if it was possible to reduce the cost of verifying computational models of tidal waves and currents. Statistical techniques were used to determine the least number of samples required, in a given situation, to remain statistically significant, and thereby reduce overall project costs. Commercial, academic, and Federal agencies could benefit by applying these techniques, without the need to 'touch' every item in the population. For example, the requirement of this project was to measure the heights and times of high and low tides at 8,000 locations for verification of computational models of tidal waves and currents. The application of the statistical techniques began with observations to determine the correctness of submitted measurement data, followed by some assumptions based on the observations. Among the assumptions were that the data were representative of data-collection techniques used at the measurement locations, that time measurements could be ignored (that is, height measurements alone would suffice), and that the height measurements were from a statistically normal distribution. Sample means and standard deviations were determined for all locations. Interval limits were determined for confidence levels of 95, 98, and 99 percent. It was found that the numbers of measurement locations needed to attain these confidence levels were 55, 78, and 96, respectively.

Source record↗

NASA NTRS1961

Application of Monte Carlo Technique for Determining Maneuvering Loads from Statistical Information on Airplane Motions

Results of a statistical analysis of horizontal-tail loads on a fighter airplane are presented. The data were obtained from a number of operational training missions with flight at altitudes up to about 50,000 feet and at Mach numbers up to 1.22. The analysis was performed to determine the feasibility of calculating horizontal-tail load from data on the flight conditions and airplane motions. In the analysis the calculated loads are compared with the measured loads for the different types of missions performed. The loads were calculated by two methods: a direct approach and a Monte Carlo technique. The procedures used and some of the problems associated with the data analysis are discussed. frequencies of occurrence of tail loads of given magnitudes are derived from statistical information on the flight quantities. In the direct method, a time history of tail load is calculated from time-history measurements of the flight quantities. The Monte Carlo method could be useful for extending loads information for design of prospective airplanes . For the Monte Carlo method, the The results indicate that the accuracy of loads, regardless of the method used for calculation, is largely dependent on the knowledge of the pertinent airplane aerodynamic characteristics and center-of-gravity location. In addition, reliable Monte Carlo results require an adequate sample of statistical data and a knowledge of the more important statistical dependencies between the various flight conditions and airplane motions.

Hamer, Harold A.↗

NASA NTRS2004

Automatic Generation of Algorithms for the Statistical Analysis of Planetary Nebulae Images

Analyzing data sets collected in experiments or by observations is a Core scientific activity. Typically, experimentd and observational data are &aught with uncertainty, and the analysis is based on a statistical model of the conjectured underlying processes, The large data volumes collected by modern instruments make computer support indispensible for this. Consequently, scientists spend significant amounts of their time with the development and refinement of the data analysis programs. AutoBayes [GF+02, FS03] is a fully automatic synthesis system for generating statistical data analysis programs. Externally, it looks like a compiler: it takes an abstract problem specification and translates it into executable code. Its input is a concise description of a data analysis problem in the form of a statistical model as shown in Figure 1; its output is optimized and fully documented C/C++ code which can be linked dynamically into the Matlab and Octave environments. Internally, however, it is quite different: AutoBayes derives a customized algorithm implementing the given model using a schema-based process, and then further refines and optimizes the algorithm into code. A schema is a parameterized code template with associated semantic constraints which define and restrict the template s applicability. The schema parameters are instantiated in a problem-specific way during synthesis as AutoBayes checks the constraints against the original model or, recursively, against emerging sub-problems. AutoBayes schema library contains problem decomposition operators (which are justified by theorems in a formal logic in the domain of Bayesian networks) as well as machine learning algorithms (e.g., EM, k-Means) and nu- meric optimization methods (e.g., Nelder-Mead simplex, conjugate gradient). AutoBayes augments this schema-based approach by symbolic computation to derive closed-form solutions whenever possible. This is a major advantage over other statistical data analysis systems which use numerical approximations even in cases where closed-form solutions exist. AutoBayes is implemented in Prolog and comprises approximately 75.000 lines of code. In this paper, we take one typical scientific data analysis problem-analyzing planetary nebulae images taken by the Hubble Space Telescope-and show how AutoBayes can be used to automate the implementation of the necessary anal- ysis programs. We initially follow the analysis described by Knuth and Hajian [KHO2] and use AutoBayes to derive code for the published models. We show the details of the code derivation process, including the symbolic computations and automatic integration of library procedures, and compare the results of the automatically generated and manually implemented code. We then go beyond the original analysis and use AutoBayes to derive code for a simple image segmentation procedure based on a mixture model which can be used to automate a manual preproceesing step. Finally, we combine the original approach with the simple segmentation which yields a more detailed analysis. This also demonstrates that AutoBayes makes it easy to combine different aspects of data analysis.

Fischer, Bernd↗

NASA NTRS2006

Calculating Statistical Orbit Distributions Using GEO Optical Observations with the Michigan Orbital Debris Survey Telescope (MODEST)

NASA's Orbital Debris measurements program has a goal to characterize the small debris environment in the geosynchronous Earth-orbit (GEO) region using optical telescopes ("small" refers to objects too small to catalog and track with current systems). Traditionally, observations of GEO and near-GEO objects involve following the object with the telescope long enough to obtain an orbit suitable for tracking purposes. Telescopes operating in survey mode, however, randomly observe objects that pass through their field of view. Typically, these short-arc observation are inadequate to obtain detailed orbits, but can be used to estimate approximate circular orbit elements (semimajor axis, inclination, and ascending node). From this information, it should be possible to make statistical inferences about the orbital distributions of the GEO population bright enough to be observed by the system. The Michigan Orbital Debris Survey Telescope (MODEST) has been making such statistical surveys of the GEO region for four years. During that time, the telescope has made enough observations in enough areas of the GEO belt to have had nearly complete coverage. That means that almost all objects in all possible orbits in the GEO and near- GEO region had a non-zero chance of being observed. Some regions (such as those near zero inclination) have had good coverage, while others are poorly covered. Nevertheless, it is possible to remove these statistical biases and reconstruct the orbit populations within the limits of sampling error. In this paper, these statistical techniques and assumptions are described, and the techniques are applied to the current MODEST data set to arrive at our best estimate of the GEO orbit population distribution.

Matney, M.↗

NASA NTRS2006

Exploring the Connection Between Sampling Problems in Bayesian Inference and Statistical Mechanics

The Bayesian and statistical mechanical communities often share the same objective in their work - estimating and integrating probability distribution functions (pdfs) describing stochastic systems, models or processes. Frequently, these pdfs are complex functions of random variables exhibiting multiple, well separated local minima. Conventional strategies for sampling such pdfs are inefficient, sometimes leading to an apparent non-ergodic behavior. Several recently developed techniques for handling this problem have been successfully applied in statistical mechanics. In the multicanonical and Wang-Landau Monte Carlo (MC) methods, the correct pdfs are recovered from uniform sampling of the parameter space by iteratively establishing proper weighting factors connecting these distributions. Trivial generalizations allow for sampling from any chosen pdf. The closely related transition matrix method relies on estimating transition probabilities between different states. All these methods proved to generate estimates of pdfs with high statistical accuracy. In another MC technique, parallel tempering, several random walks, each corresponding to a different value of a parameter (e.g. "temperature"), are generated and occasionally exchanged using the Metropolis criterion. This method can be considered as a statistically correct version of simulated annealing. An alternative approach is to represent the set of independent variables as a Hamiltonian system. Considerab!e progress has been made in understanding how to ensure that the system obeys the equipartition theorem or, equivalently, that coupling between the variables is correctly described. Then a host of techniques developed for dynamical systems can be used. Among them, probably the most powerful is the Adaptive Biasing Force method, in which thermodynamic integration and biased sampling are combined to yield very efficient estimates of pdfs. The third class of methods deals with transitions between states described by rate constants. These problems are isomorphic with chemical kinetics problems. Recently, several efficient techniques for this purpose have been developed based on the approach originally proposed by Gillespie. Although the utility of the techniques mentioned above for Bayesian problems has not been determined, further research along these lines is warranted

Pohorille, Andrew↗

NASA NTRS1997

Statistical Modeling of Landscape Pattern and the Effects of Coarse Spatial Resolution

Analysis of classified satellite imagery was conducted to characterize errors in estimates of area based on coarse resolution satellite imagery which are due to distortions in sizes of small fragments, and to explore the feasibility of correcting for these errors using a statistical modeling approach. Sizes of bodies of open water on European Remote-Sensing (ERS-1) synthetic aperture radar (SAR) and fire scars on Landsat Multispectral Scanner (MSS) imagery were measured using geographic information system (GIS) software. Statistical analysis of the smaller scars and ponds as observed with this imagery of relatively fine resolution demonstrated that the distribution of the sizes could be modeled by either of two types of statistical distributions-a power distribution related to fractal processes or a simple exponential distribution. A comparison of the size distribution of small burn scars as observed with Landsat with the distribution observed with the much coarser Advanced Very High Resolution Radiometer (AVHRR) imagery revealed effects of coarse resolution, each of which were associated with a particular size range, and whose net effect accounted for the bias in total burn area due to coarse spatial resolution. Size dependent area correction procedures are therefore feasible. A statistical modeling approach to area estimation, based models of the underlying and observed size distributions of fragments in the scene, may lead to a procedure which requires little or no ancillary data.

Hlavka, C. A.↗

NASA NTRS2010

Weighting Statistical Inputs for Data Used to Support Effective Decision Making During Severe Emergency Weather and Environmental Events

National Aeronautical and Space Administration (NASA) weather and atmospheric environmental organizations are insatiable consumers of geophysical, hydrometeorological and solar weather statistics. The expanding array of internet-worked sensors producing targeted physical measurements has generated an almost factorial explosion of near real-time inputs to topical statistical datasets. Normalizing and value-based parsing of such statistical datasets in support of time-constrained weather and environmental alerts and warnings is essential, even with dedicated high-performance computational capabilities. What are the optimal indicators for advanced decision making? How do we recognize the line between sufficient statistical sampling and excessive, mission destructive sampling ? How do we assure that the normalization and parsing process, when interpolated through numerical models, yields accurate and actionable alerts and warnings? This presentation will address the integrated means and methods to achieve desired outputs for NASA and consumers of its data.

Gardner, Adrian↗

NASA NTRS2011

Characterizing the Joint Effect of Diverse Test-Statistic Correlation Structures and Effect Size on False Discovery Rates in a Multiple-Comparison Study of Many Outcome Measures

In their 2009 Annals of Statistics paper, Gavrilov, Benjamini, and Sarkar report the results of a simulation assessing the robustness of their adaptive step-down procedure (GBS) for controlling the false discovery rate (FDR) when normally distributed test statistics are serially correlated. In this study we extend the investigation to the case of multiple comparisons involving correlated non-central t-statistics, in particular when several treatments or time periods are being compared to a control in a repeated-measures design with many dependent outcome measures. In addition, we consider several dependence structures other than serial correlation and illustrate how the FDR depends on the interaction between effect size and the type of correlation structure as indexed by Foerstner s distance metric from an identity. The relationship between the correlation matrix R of the original dependent variables and R~, the correlation matrix of associated t-statistics is also studied. In general R~ depends not only on R, but also on sample size and the signed effect sizes for the multiple comparisons.

Feiveson, Alan H.↗

NASA NTRS2011

Ten Years of Cloud Properties from MODIS: Global Statistics and Use in Climate Model Evaluation

The NASA Moderate Resolution Imaging Spectroradiometer (MODIS), launched onboard the Terra and Aqua spacecrafts, began Earth observations on February 24, 2000 and June 24,2002, respectively. Among the algorithms developed and applied to this sensor, a suite of cloud products includes cloud masking/detection, cloud-top properties (temperature, pressure), and optical properties (optical thickness, effective particle radius, water path, and thermodynamic phase). All cloud algorithms underwent numerous changes and enhancements between for the latest Collection 5 production version; this process continues with the current Collection 6 development. We will show example MODIS Collection 5 cloud climatologies derived from global spatial . and temporal aggregations provided in the archived gridded Level-3 MODIS atmosphere team product (product names MOD08 and MYD08 for MODIS Terra and Aqua, respectively). Data sets in this Level-3 product include scalar statistics as well as 1- and 2-D histograms of many cloud properties, allowing for higher order information and correlation studies. In addition to these statistics, we will show trends and statistical significance in annual and seasonal means for a variety of the MODIS cloud properties, as well as the time required for detection given assumed trends. To assist in climate model evaluation, we have developed a MODIS cloud simulator with an accompanying netCDF file containing subsetted monthly Level-3 statistical data sets that correspond to the simulator output. Correlations of cloud properties with ENSO offer the potential to evaluate model cloud sensitivity; initial results will be discussed.

Platnick, Steven E.↗

NASA NTRS2005

Tools for Basic Statistical Analysis

Statistical Analysis Toolset is a collection of eight Microsoft Excel spreadsheet programs, each of which performs calculations pertaining to an aspect of statistical analysis. These programs present input and output data in user-friendly, menu-driven formats, with automatic execution. The following types of calculations are performed: Descriptive statistics are computed for a set of data x(i) (i = 1, 2, 3 . . . ) entered by the user. Normal Distribution Estimates will calculate the statistical value that corresponds to cumulative probability values, given a sample mean and standard deviation of the normal distribution. Normal Distribution from two Data Points will extend and generate a cumulative normal distribution for the user, given two data points and their associated probability values. Two programs perform two-way analysis of variance (ANOVA) with no replication or generalized ANOVA for two factors with four levels and three repetitions. Linear Regression-ANOVA will curvefit data to the linear equation y=f(x) and will do an ANOVA to check its significance.

Luz, Paul L.↗

NASA NTRS2011

Estimated Accuracy of Three Common Trajectory Statistical Methods

Three well-known trajectory statistical methods (TSMs), namely concentration field (CF), concentration weighted trajectory (CWT), and potential source contribution function (PSCF) methods were tested using known sources and artificially generated data sets to determine the ability of TSMs to reproduce spatial distribution of the sources. In the works by other authors, the accuracy of the trajectory statistical methods was estimated for particular species and at specified receptor locations. We have obtained a more general statistical estimation of the accuracy of source reconstruction and have found optimum conditions to reconstruct source distributions of atmospheric trace substances. Only virtual pollutants of the primary type were considered. In real world experiments, TSMs are intended for application to a priori unknown sources. Therefore, the accuracy of TSMs has to be tested with all possible spatial distributions of sources. An ensemble of geographical distributions of virtual sources was generated. Spearman s rank order correlation coefficient between spatial distributions of the known virtual and the reconstructed sources was taken to be a quantitative measure of the accuracy. Statistical estimates of the mean correlation coefficient and a range of the most probable values of correlation coefficients were obtained. All the TSMs that were considered here showed similar close results. The maximum of the ratio of the mean correlation to the width of the correlation interval containing the most probable correlation values determines the optimum conditions for reconstruction. An optimal geographical domain roughly coincides with the area supplying most of the substance to the receptor. The optimal domain s size is dependent on the substance decay time. Under optimum reconstruction conditions, the mean correlation coefficients can reach 0.70 0.75. The boundaries of the interval with the most probable correlation values are 0.6 0.9 for the decay time of 240 h and 0.5 0.95 for the decay time of 12 h. The best results of source reconstruction can be expected for the trace substances with a decay time on the order of several days. Although the methods considered in this paper do not guarantee high accuracy they are computationally simple and fast. Using the TSMs in optimum conditions and taking into account the range of uncertainties, one can obtain a first hint on potential source areas.

Kabashnikov, Vitaliy P.↗

NASA NTRS2011

Statistically Optimized Inversion Algorithm for Enhanced Retrieval of Aerosol Properties from Spectral Multi-Angle Polarimetric Satellite Observations

The proposed development is an attempt to enhance aerosol retrieval by emphasizing statistical optimization in inversion of advanced satellite observations. This optimization concept improves retrieval accuracy relying on the knowledge of measurement error distribution. Efficient application of such optimization requires pronounced data redundancy (excess of the measurements number over number of unknowns) that is not common in satellite observations. The POLDER imager on board the PARASOL microsatellite registers spectral polarimetric characteristics of the reflected atmospheric radiation at up to 16 viewing directions over each observed pixel. The completeness of such observations is notably higher than for most currently operating passive satellite aerosol sensors. This provides an opportunity for profound utilization of statistical optimization principles in satellite data inversion. The proposed retrieval scheme is designed as statistically optimized multi-variable fitting of all available angular observations obtained by the POLDER sensor in the window spectral channels where absorption by gas is minimal. The total number of such observations by PARASOL always exceeds a hundred over each pixel and the statistical optimization concept promises to be efficient even if the algorithm retrieves several tens of aerosol parameters. Based on this idea, the proposed algorithm uses a large number of unknowns and is aimed at retrieval of extended set of parameters affecting measured radiation.

Dubovik, O↗

NASA NTRS2013

Spatial Statistical Data Fusion (SSDF)

As remote sensing for scientific purposes has transitioned from an experimental technology to an operational one, the selection of instruments has become more coordinated, so that the scientific community can exploit complementary measurements. However, tech nological and scientific heterogeneity across devices means that the statistical characteristics of the data they collect are different. The challenge addressed here is how to combine heterogeneous remote sensing data sets in a way that yields optimal statistical estimates of the underlying geophysical field, and provides rigorous uncertainty measures for those estimates. Different remote sensing data sets may have different spatial resolutions, different measurement error biases and variances, and other disparate characteristics. A state-of-the-art spatial statistical model was used to relate the true, but not directly observed, geophysical field to noisy, spatial aggregates observed by remote sensing instruments. The spatial covariances of the true field and the covariances of the true field with the observations were modeled. The observations are spatial averages of the true field values, over pixels, with different measurement noise superimposed. A kriging framework is used to infer optimal (minimum mean squared error and unbiased) estimates of the true field at point locations from pixel-level, noisy observations. A key feature of the spatial statistical model is the spatial mixed effects model that underlies it. The approach models the spatial covariance function of the underlying field using linear combinations of basis functions of fixed size. Approaches based on kriging require the inversion of very large spatial covariance matrices, and this is usually done by making simplifying assumptions about spatial covariance structure that simply do not hold for geophysical variables. In contrast, this method does not require these assumptions, and is also computationally much faster. This method is fundamentally different than other approaches to data fusion for remote sensing data because it is inferential rather than merely descriptive. All approaches combine data in a way that minimizes some specified loss function. Most of these are more or less ad hoc criteria based on what looks good to the eye, or some criteria that relate only to the data at hand.

Braverman, Amy J.↗

NASA NTRS2013

Statistical Downscaling and Bias Correction of Climate Model Outputs for Climate Change Impact Assessment in the U.S. Northeast

Statistical downscaling can be used to efficiently downscale a large number of General Circulation Model (GCM) outputs to a fine temporal and spatial scale. To facilitate regional impact assessments, this study statistically downscales (to 1∕8deg spatial resolution) and corrects the bias of daily maximum and minimum temperature and daily precipitation data from six GCMs and four Regional Climate Models (RCMs) for the northeast United States (US) using the Statistical Downscaling and Bias Correction (SDBC) approach. Based on these downscaled data from multiple models, five extreme indices were analyzed for the future climate to quantify future changes of climate extremes. For a subset of models and indices, results based on raw and bias corrected model outputs for the present-day climate were compared with observations, which demonstrated that bias correction is important not only for GCM outputs, but also for RCM outputs. For future climate, bias correction led to a higher level of agreements among the models in predicting the magnitude and capturing the spatial pattern of the extreme climate indices. We found that the incorporation of dynamical downscaling as an intermediate step does not lead to considerable differences in the results of statistical downscaling for the study domain.

extreme climate index↗

NASA NTRS2014

Impact of Satellite Viewing-Swath Width on Global and Regional Aerosol Optical Thickness Statistics and Trends

We use the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite aerosol optical thickness (AOT) product to assess the impact of reduced swath width on global and regional AOT statistics and trends. Alongtrack and across-track sampling strategies are employed, in which the full MODIS data set is sub-sampled with various narrow-swath (approximately 400-800 km) and single pixel width (approximately 10 km) configurations. Although view-angle artifacts in the MODIS AOT retrieval confound direct comparisons between averages derived from different sub-samples, careful analysis shows that with many portions of the Earth essentially unobserved, spatial sampling introduces uncertainty in the derived seasonal-regional mean AOT. These AOT spatial sampling artifacts comprise up to 60%of the full-swath AOT value under moderate aerosol loading, and can be as large as 0.1 in some regions under high aerosol loading. Compared to full-swath observations, narrower swath and single pixel width sampling exhibits a reduced ability to detect AOT trends with statistical significance. On the other hand, estimates of the global, annual mean AOT do not vary significantly from the full-swath values as spatial sampling is reduced. Aggregation of the MODIS data at coarse grid scales (10 deg) shows consistency in the aerosol trends across sampling strategies, with increased statistical confidence, but quantitative errors in the derived trends are found even for the full-swath data when compared to high spatial resolution (0.5 deg) aggregations. Using results of a model-derived aerosol reanalysis, we find consistency in our conclusions about a seasonal-regional spatial sampling artifact in AOT Furthermore, the model shows that reduced spatial sampling can amount to uncertainty in computed shortwave top-ofatmosphere aerosol radiative forcing of 2-3 W m(sup−2). These artifacts are lower bounds, as possibly other unconsidered sampling strategies would perform less well. These results suggest that future aerosol satellite missions having significantly less than full-swath viewing are unlikely to sample the true AOT distribution well enough to obtain the statistics needed to reduce uncertainty in aerosol direct forcing of climate.

remote sensing↗

NASA NTRS2015

Statistical Analysis of Large Simulated Yield Datasets for Studying Climate Effects

Many studies have been carried out during the last decade to study the effect of climate change on crop yields and other key crop characteristics. In these studies, one or several crop models were used to simulate crop growth and development for different climate scenarios that correspond to different projections of atmospheric CO2 concentration, temperature, and rainfall changes (Semenov et al., 1996; Tubiello and Ewert, 2002; White et al., 2011). The Agricultural Model Intercomparison and Improvement Project (AgMIP; Rosenzweig et al., 2013) builds on these studies with the goal of using an ensemble of multiple crop models in order to assess effects of climate change scenarios for several crops in contrasting environments. These studies generate large datasets, including thousands of simulated crop yield data. They include series of yield values obtained by combining several crop models with different climate scenarios that are defined by several climatic variables (temperature, CO2, rainfall, etc.). Such datasets potentially provide useful information on the possible effects of different climate change scenarios on crop yields. However, it is sometimes difficult to analyze these datasets and to summarize them in a useful way due to their structural complexity; simulated yield data can differ among contrasting climate scenarios, sites, and crop models. Another issue is that it is not straightforward to extrapolate the results obtained for the scenarios to alternative climate change scenarios not initially included in the simulation protocols. Additional dynamic crop model simulations for new climate change scenarios are an option but this approach is costly, especially when a large number of crop models are used to generate the simulated data, as in AgMIP. Statistical models have been used to analyze responses of measured yield data to climate variables in past studies (Lobell et al., 2011), but the use of a statistical model to analyze yields simulated by complex process-based crop models is a rather new idea. We demonstrate herewith that statistical methods can play an important role in analyzing simulated yield data sets obtained from the ensembles of process-based crop models. Formal statistical analysis is helpful to estimate the effects of different climatic variables on yield, and to describe the between-model variability of these effects.

climate↗