Bayesian Inference in Large-scale Problems

Many modern applications fall into the category of "large-scale" statistical problems, in which both the number of observations n and the number of features or parameters p may be large. Many existing methods focus on point estimation, despite the continued relevance of uncertainty quantification in the sciences, where the number of parameters to estimate often exceeds the sample size, despite huge increases in the value of n typically seen in many fields. Thus, the tendency in some areas of industry to dispense with traditional statistical analysis on the basis that "n=all" is of little relevance outside of certain narrow applications. The main result of the Big Data revolution in most fields has instead been to make computation much harder without reducing the importance of uncertainty quantification. Bayesian methods excel at uncertainty quantification, but often scale poorly relative to alternatives. This conflict between the statistical advantages of Bayesian procedures and their substantial computational disadvantages is perhaps the greatest challenge facing modern Bayesian statistics, and is the primary motivation for the work presented here.

Two general strategies for scaling Bayesian inference are considered. The first is the development of methods that lend themselves to faster computation, and the second is design and characterization of computational algorithms that scale better in n or p. In the first instance, the focus is on joint inference outside of the standard problem of multivariate continuous data that has been a major focus of previous theoretical work in this area. In the second area, we pursue strategies for improving the speed of Markov chain Monte Carlo algorithms, and characterizing their performance in large-scale settings. Throughout, the focus is on rigorous theoretical evaluation combined with empirical demonstrations of performance and concordance with the theory.

One topic we consider is modeling the joint distribution of multivariate categorical data, often summarized in a contingency table. Contingency table analysis routinely relies on log-linear models, with latent structure analysis providing a common alternative. Latent structure models lead to a reduced rank tensor factorization of the probability mass function for multivariate categorical data, while log-linear models achieve dimensionality reduction through sparsity. Little is known about the relationship between these notions of dimensionality reduction in the two paradigms. In Chapter 2, we derive several results relating the support of a log-linear model to nonnegative ranks of the associated probability tensor. Motivated by these findings, we propose a new collapsed Tucker class of tensor decompositions, which bridge existing PARAFAC and Tucker decompositions, providing a more flexible framework for parsimoniously characterizing multivariate categorical data. Taking a Bayesian approach to inference, we illustrate empirical advantages of the new decompositions.

Latent class models for the joint distribution of multivariate categorical, such as the PARAFAC decomposition, data play an important role in the analysis of population structure. In this context, the number of latent classes is interpreted as the number of genetically distinct subpopulations of an organism, an important factor in the analysis of evolutionary processes and conservation status. Existing methods focus on point estimates of the number of subpopulations, and lack robust uncertainty quantification. Moreover, whether the number of latent classes in these models is even an identified parameter is an open question. In Chapter 3, we show that when the model is properly specified, the correct number of subpopulations can be recovered almost surely. We then propose an alternative method for estimating the number of latent subpopulations that provides good quantification of uncertainty, and provide a simple procedure for verifying that the proposed method is consistent for the number of subpopulations. The performance of the model in estimating the number of subpopulations and other common population structure inference problems is assessed in simulations and a real data application.

In contingency table analysis, sparse data is frequently encountered for even modest numbers of variables, resulting in non-existence of maximum likelihood estimates. A common solution is to obtain regularized estimates of the parameters of a log-linear model. Bayesian methods provide a coherent approach to regularization, but are often computationally intensive. Conjugate priors ease computational demands, but the conjugate Diaconis--Ylvisaker priors for the parameters of log-linear models do not give rise to closed form credible regions, complicating posterior inference. In Chapter 4 we derive the optimal Gaussian approximation to the posterior for log-linear models with Diaconis--Ylvisaker priors, and provide convergence rate and finite-sample bounds for the Kullback-Leibler divergence between the exact posterior and the optimal Gaussian approximation. We demonstrate empirically in simulations and a real data application that the approximation is highly accurate, even in relatively small samples. The proposed approximation provides a computationally scalable and principled approach to regularized estimation and approximate Bayesian inference for log-linear models.

Another challenging and somewhat non-standard joint modeling problem is inference on tail dependence in stochastic processes. In applications where extreme dependence is of interest, data are almost always time-indexed. Existing methods for inference and modeling in this setting often cluster extreme events or choose window sizes with the goal of preserving temporal information. In Chapter 5, we propose an alternative paradigm for inference on tail dependence in stochastic processes with arbitrary temporal dependence structure in the extremes, based on the idea that the information on strength of tail dependence and the temporal structure in this dependence are both encoded in waiting times between exceedances of high thresholds. We construct a class of time-indexed stochastic processes with tail dependence obtained by endowing the support points in de Haan's spectral representation of max-stable processes with velocities and lifetimes. We extend Smith's model to these max-stable velocity processes and obtain the distribution of waiting times between extreme events at multiple locations. Motivated by this result, a new definition of tail dependence is proposed that is a function of the distribution of waiting times between threshold exceedances, and an inferential framework is constructed for estimating the strength of extremal dependence and quantifying uncertainty in this paradigm. The method is applied to climatological, financial, and electrophysiology data.

The remainder of this thesis focuses on posterior computation by Markov chain Monte Carlo. The Markov Chain Monte Carlo method is the dominant paradigm for posterior computation in Bayesian analysis. It has long been common to control computation time by making approximations to the Markov transition kernel. Comparatively little attention has been paid to convergence and estimation error in these approximating Markov Chains. In Chapter 6, we propose a framework for assessing when to use approximations in MCMC algorithms, and how much error in the transition kernel should be tolerated to obtain optimal estimation performance with respect to a specified loss function and computational budget. The results require only ergodicity of the exact kernel and control of the kernel approximation accuracy. The theoretical framework is applied to approximations based on random subsets of data, low-rank approximations of Gaussian processes, and a novel approximating Markov chain for discrete mixture models.

Data augmentation Gibbs samplers are arguably the most popular class of algorithm for approximately sampling from the posterior distribution for the parameters of generalized linear models. The truncated Normal and Polya-Gamma data augmentation samplers are standard examples for probit and logit links, respectively. Motivated by an important problem in quantitative advertising, in Chapter 7 we consider the application of these algorithms to modeling rare events. We show that when the sample size is large but the observed number of successes is small, these data augmentation samplers mix very slowly, with a spectral gap that converges to zero at a rate at least proportional to the reciprocal of the square root of the sample size up to a log factor. In simulation studies, moderate sample sizes result in high autocorrelations and small effective sample sizes. Similar empirical results are observed for related data augmentation samplers for multinomial logit and probit models. When applied to a real quantitative advertising dataset, the data augmentation samplers mix very poorly. Conversely, Hamiltonian Monte Carlo and a type of independence chain Metropolis algorithm show good mixing on the same dataset.

Dynamic Programming Approaches for Estimating and Applying Large-scale Discrete Choice Models

People go through their life making all kinds of decisions, and some of these decisions affect their demand for transportation, for example, their choices of where to live and where to work, how and when to travel and which route to take. Transport related choices are typically time dependent and characterized by large number of alternatives that can be spatially correlated. This thesis deals with models that can be used to analyze and predict discrete choices in large-scale networks. The proposed models and methods are highly relevant for, but not limited to, transport applications. We model decisions as sequences of choices within the dynamic discrete choice framework, also known as parametric Markov decision processes. Such models are known to be difficult to estimate and to apply to make predictions because dynamic programming problems need to be solved in order to compute choice probabilities. In this thesis we show that it is possible to explore the network structure and the flexibility of dynamic programming so that the dynamic discrete choice modeling approach is not only useful to model time dependent choices, but also makes it easier to model large-scale static choices. The thesis consists of seven articles containing a number of models and methods for estimating, applying and testing large-scale discrete choice models. In the following we group the contributions under three themes: route choice modeling, large-scale multivariate extreme value (MEV) model estimation and nonlinear optimization algorithms. Five articles are related to route choice modeling. We propose different dynamic discrete choice models that allow paths to be correlated based on the MEV and mixed logit models. The resulting route choice models become expensive to estimate and we deal with this challenge by proposing innovative methods that allow to reduce the estimation cost. For example, we propose a decomposition method that not only opens up for possibility of mixing, but also speeds up the estimation for simple logit models, which has implications also for traffic simulation. Moreover, we compare the utility maximization and regret minimization decision rules, and we propose a misspecification test for logit-based route choice models. The second theme is related to the estimation of static discrete choice models with large choice sets. We establish that a class of MEV models can be reformulated as dynamic discrete choice models on the networks of correlation structures. These dynamic models can then be estimated quickly using dynamic programming techniques and an efficient nonlinear optimization algorithm. Finally, the third theme focuses on structured quasi-Newton techniques for estimating discrete choice models by maximum likelihood. We examine and adapt switching methods that can be easily integrated into usual optimization algorithms (line search and trust region) to accelerate the estimation process. The proposed dynamic discrete choice models and estimation methods can be used in various discrete choice applications. In the area of big data analytics, models that can deal with large choice sets and sequential choices are important. Our research can therefore be of interest in various demand analysis applications (predictive analytics) or can be integrated with optimization models (prescriptive analytics). Furthermore, our studies indicate the potential of dynamic programming techniques in this context, even for static models, which opens up a variety of future research directions.

Modeling the impact of large-scale solid mass transfers on the gravity field

The focus of this thesis is to explore and quantify the response of large-scale solid mass transfer events on satellite-based gravity observations. The gravity signature of large-scale solid mass transfers has not been deeply explored yet; mainly due to the lack of significant events during dedicated satellite gravity missions‘ lifespans. In light of the next generation of gravity missions, the feasibility of employing satellite gravity observations to detect submarine and surface mass transfers is of importance for geoscience (improves the understanding of geodynamic processes) and for geodesy (improves the understanding of the dynamic gravity field). The aim of this thesis is twofold and focuses on assessing the feasibility of using satellite gravity observations for detecting large-scale solid mass transfers and on modeling the impact on the gravity field caused by these events. A methodology that employs 3D forward modeling simulations and 2D wavelet multiresolution analysis is suggested to estimate the impact of solid mass transfers on satellite gravity observations. The gravity signature of various submarine and subaerial events that occurred in the past was estimated. Case studies were conducted to assess the sensitivity and resolvability required in order to observe gravity differences caused by solid mass transfers. Simulation studies were also employed in order to assess the expected contribution of the Next Generation of Gravity Missions for this application.

A Scalable Design Framework for Variability Management in Large-Scale Software Product Lines

Variability management is one of the major challenges in software product line adoption, since it needs to be efficiently managed at various levels of the software product line development process (e.g., requirement analysis, design, implementation, etc.). One of the main challenges within variability management is the handling and effective visualization of large-scale (industry-size) models, which in many projects, can reach the order of thousands, along with the dependency relationships that exist among them. These have raised many concerns regarding the scalability of current variability management tools and techniques and their lack of industrial adoption. To address the scalability issues, this work employed a combination of quantitative and qualitative research methods to identify the reasons behind the limited scalability of existing variability management tools and techniques. In addition to producing a comprehensive catalogue of existing tools, the outcome form this stage helped understand the major limitations of existing tools. Based on the findings, a novel approach was created for managing variability that employed two main principles for supporting scalability. First, the separation-of-concerns principle was employed by creating multiple views of variability models to alleviate information overload. Second, hyperbolic trees were used to visualise models (compared to Euclidian space trees traditionally used). The result was an approach that can represent models encompassing hundreds of variability points and complex relationships. These concepts were demonstrated by implementing them in an existing variability management tool and using it to model a real-life product line with over a thousand variability points. Finally, in order to assess the work, an evaluation framework was designed based on various established usability assessment best practices and standards. The framework was then used with several case studies to benchmark the performance of this work against other existing tools.

Dynamic Programming Approaches for Estimating and Applying Large-scale Discrete Choice Models

People go through their life making all kinds of decisions, and some of these decisions affect their demand for transportation, for example, their choices of where to live and where to work, how and when to travel and which route to take. Transport related choices are typically time dependent and characterized by large number of alternatives that can be spatially correlated. This thesis deals with models that can be used to analyze and predict discrete choices in large-scale networks. The proposed models and methods are highly relevant for, but not limited to, transport applications. We model decisions as sequences of choices within the dynamic discrete choice framework, also known as parametric Markov decision processes. Such models are known to be difficult to estimate and to apply to make predictions because dynamic programming problems need to be solved in order to compute choice probabilities. In this thesis we show that it is possible to explore the network structure and the flexibility of dynamic programming so that the dynamic discrete choice modeling approach is not only useful to model time dependent choices, but also makes it easier to model large-scale static choices. The thesis consists of seven articles containing a number of models and methods for estimating, applying and testing large-scale discrete choice models. In the following we group the contributions under three themes: route choice modeling, large-scale multivariate extreme value (MEV) model estimation and nonlinear optimization algorithms. Five articles are related to route choice modeling. We propose different dynamic discrete choice models that allow paths to be correlated based on the MEV and mixed logit models. The resulting route choice models become expensive to estimate and we deal with this challenge by proposing innovative methods that allow to reduce the estimation cost. For example, we propose a decomposition method that not only opens up for possibility of mixing, but also speeds up the estimation for simple logit models, which has implications also for traffic simulation. Moreover, we compare the utility maximization and regret minimization decision rules, and we propose a misspecification test for logit-based route choice models. The second theme is related to the estimation of static discrete choice models with large choice sets. We establish that a class of MEV models can be reformulated as dynamic discrete choice models on the networks of correlation structures. These dynamic models can then be estimated quickly using dynamic programming techniques and an efficient nonlinear optimization algorithm. Finally, the third theme focuses on structured quasi-Newton techniques for estimating discrete choice models by maximum likelihood. We examine and adapt switching methods that can be easily integrated into usual optimization algorithms (line search and trust region) to accelerate the estimation process. The proposed dynamic discrete choice models and estimation methods can be used in various discrete choice applications. In the area of big data analytics, models that can deal with large choice sets and sequential choices are important. Our research can therefore be of interest in various demand analysis applications (predictive analytics) or can be integrated with optimization models (prescriptive analytics). Furthermore, our studies indicate the potential of dynamic programming techniques in this context, even for static models, which opens up a variety of future research directions.

Large-scale neural dynamics: simple and complex

We review the use of neural field models for modelling the brain at the large scales necessary for interpreting EEG, fMRI, MEG and optical imaging data. Albeit a framework that is limited to coarse-grained or mean-field activity, neural field models provide a framework for unifying data from different imaging modalities. Starting with a description of neural mass models we build to spatially extended cortical models of layered two-dimensional sheets with long range axonal connections mediating synaptic interactions. Reformulations of the fundamental non-local mathematical model in terms of more familiar local differential (brain wave) equations are described. Techniques for the analysis of such models, including how to determine the onset of spatio-temporal pattern forming instabilities, are reviewed. Extensions of the basic formalism to treat refractoriness, adaptive feedback and inhomogeneous connectivity are described along with open challenges for the development of multi-scale models that can integrate macroscopic models at large spatial scales with models at the microscopic scale.

Large-scale screening of a targeted Enterococcus faecalis mutant library identifies envelope fitness factors

Spread of antibiotic resistance among bacteria responsible for nosocomial and community-acquired infections urges for novel therapeutic or prophylactic targets and for innovative pathogen-specific antibacterial compounds. Major challenges are posed by opportunistic pathogens belonging to the low GC% gram-positive bacteria. Among those, Enterococcus faecalis is a leading cause of hospital-acquired infections associated with life-threatening issues and increased hospital costs. To better understand the molecular properties of enterococci that may be required for virulence, and that may explain the emergence of these bacteria in nosocomial infections, we performed the first large-scale functional analysis of E. faecalis V583, the first vancomycin-resistant isolate from a human bloodstream infection. E. faecalis V583 is within the high-risk clonal complex 2 group, which comprises mostly isolates derived from hospital infections worldwide. We conducted broad-range screenings of candidate genes likely involved in host adaptation (e.g., colonization and/or virulence). For this purpose, a library was constructed of targeted insertion mutations in 177 genes encoding putative surface or stress-response factors. Individual mutants were subsequently tested for their i) resistance to oxidative stress, ii) antibiotic resistance, iii) resistance to opsonophagocytosis, iv) adherence to the human colon carcinoma Caco-2 epithelial cells and v) virulence in a surrogate insect model. Our results identified a number of factors that are involved in the interaction between enterococci and their host environments. Their predicted functions highlight the importance of cell envelope glycopolymers in E. faecalis host adaptation. This study provides a valuable genetic database for understanding the steps leading E. faecalis to opportunistic virulence.

Time-frequency and time-scale analysis of Barkhausen noise signals

Carrying out information about the microstructure and stress behaviour of ferromagnetic steels, magnetic Barkhausen noise (MBN) has been used as a basis for effective non-destructive testing methods, opening new areas in industrial applications. One of the factors that determines the quality and reliability of the MBN analysis is the way information is extracted from the signal. Commonly, simple scalar parameters are used to characterize the information content, such as amplitude maxima and signal root mean square. This paper presents a new approach based on the time-frequency analysis. The experimental test case relates the use of MBN signals to characterize hardness gradients in a AISI4140 steel. To that purpose different time-frequency (TFR) and time-scale (TSR) representations such as the spectrogram, the Wigner-Ville distribution, the Capongram, the ARgram obtained from an AutoRegressive model, the scalogram, and the Mellingram obtained from a Mellin transform are assessed. It is shown that, due to nonstationary characteristics of the MBN, TFRs can provide a rich and new panorama of these signals. Extraction techniques of some time-frequency parameters are used to allow a diagnostic process. Comparison with results obtained by the classical method highlights the improvement on the diagnosis provided by the method proposed.

Regional scale analysis of landform configuration with base-level (isobase) maps

Base-level maps (or ""isobase maps"", as originally defined by Filosofov, 1960), express a relationship between valley order and topography. The base-level map can be seen as a ""simplified"" version of the original topographic surface, from which the ""noise"" of the low-order stream erosion was removed. This method is able to identify areas with possible tectonic influence even within lithologically uniform domains. Base-level maps have been recently applied in semi-detail scale (e.g., 1:50 000 or larger) morphotectonic analysis. In this paper, we present an evaluation of the method's applicability in regional-scale analysis (e.g., 1:250 000 or smaller). A test area was selected in northern Brazil, at the lower course of the Araguaia and Tocantins rivers. The drainage network extracted from SRTM30_PLUS DEMs with spatial resolution of approximately 900 m was visually compared with available topographic maps and considered to be compatible with a 1:1,000 000 scale. Regarding the interpretation of regional-scale morphostructures, the map constructed with 2nd and 3rd-order valleys was considered to present the best results. Some of the interpreted base-level anomalies correspond to important shear zones and geological contacts present in the 1:5 000 000 Geological Map of South America. Others have no correspondence with mapped Precambrian structures and are considered to represent younger, probably neotectonic, features. A strong E-W orientation of the base-level lines over the inflexion of the Araguaia and Tocantins rivers, suggest a major drainage capture. A N-S topographic swath profile over the Tocantins and Araguaia rivers reveals a topographic pattern which, allied with seismic data showing a roughly N-S direction of extension in the area, lead us to interpret this lineament as an E-W, southward-dipping normal fault. There is also a good visual correspondence between the base-level lineaments and geophysical anomalies. A NW-SE lineament in the southeast of the study area partially corresponds to the northern border of the Mosquito lava field, of Jurassic age, and a NW-SE lineament traced in the northeastern sector of the study area can be interpreted as the Picos-Santa Ines lineament, identifiable in geophysical maps but with little expression in hypsometric or topographic maps.

GPIomics: global analysis of glycosylphosphatidylinositol-anchored molecules of Trypanosoma cruzi

Glycosylphosphatidylinositol (GPI) anchoring is a common, relevant posttranslational modification of eukaryotic surface proteins. Here, we developed a fast, simple, and highly sensitive (high attomole-low femtomole range) method that uses liquid chromatography-tandem mass spectrometry (LC-MS(n)) for the first large-scale analysis of GPI-anchored molecules (i.e., the GPIome) of a eukaryote, Trypanosoma cruzi, the etiologic agent of Chagas disease. Our genome-wise prediction analysis revealed that approximately 12% of T. cruzi genes possibly encode GPI-anchored proteins. By analyzing the GPIome of T. cruzi insect-dwelling epimastigote stage using LC-MS(n), we identified 90 GPI species, of which 79 were novel. Moreover, we determined that mucins coded by the T. cruzi small mucin-like gene (TcSMUG S) family are the major GPI-anchored proteins expressed on the epimastigote cell surface. TcSMUG S mucin mature sequences are short (56-85 amino acids) and highly O-glycosylated, and contain few proteolytic sites, therefore, less likely susceptible to proteases of the midgut of the insect vector. We propose that our approach could be used for the high throughput GPIomic analysis of other lower and higher eukaryotes. Molecular Systems Biology 7 April 2009; doi:10.1038/msb.2009.13

Enzymatic Resolution of Racemic Sulcatol by Lipase from Candida Antarctica in a Large Scale

Large scale enzymatic resolution of racemic sulcatol 2 has been useful for stereoselective biocatalysis. This reaction was fast and selective, using vinyl acetate as donor of acyl group and lipase from Candida antarctica (CALB) as catalyst. The large scale reaction (5.0 g, 39 mmol) afforded high optical purities for S-(+)-sulcatol 2 and R-(+)-sulcatyl acetate 3, i.e., ee > 99 per cent and good yields (45 per cent) within a short time (40 min). Thermodynamic parameters for the chemoesterification of sulcatol 2 by vinyl acetate were evaluated. The enthalpy and Gibbs free energy values of this reaction were negative, indicating that this process is exothermic and spontaneous which is in agreement with the reaction obtained enzymatically.

Node-Depth Encoding and Multiobjective Evolutionary Algorithm Applied to Large-Scale Distribution System Reconfiguration

The power loss reduction in distribution systems (DSs) is a nonlinear and multiobjective problem. Service restoration in DSs is even computationally hard since it additionally requires a solution in real-time. Both DS problems are computationally complex. For large-scale networks, the usual problem formulation has thousands of constraint equations. The node-depth encoding (NDE) enables a modeling of DSs problems that eliminates several constraint equations from the usual formulation, making the problem solution simpler. On the other hand, a multiobjective evolutionary algorithm (EA) based on subpopulation tables adequately models several objectives and constraints, enabling a better exploration of the search space. The combination of the multiobjective EA with NDE (MEAN) results in the proposed approach for solving DSs problems for large-scale networks. Simulation results have shown the MEAN is able to find adequate restoration plans for a real DS with 3860 buses and 632 switches in a running time of 0.68 s. Moreover, the MEAN has shown a sublinear running time in function of the system size. Tests with networks ranging from 632 to 5166 switches indicate that the MEAN can find network configurations corresponding to a power loss reduction of 27.64% for very large networks requiring relatively low running time.

Planning and operation of large-scale water distribution systems with preemptive priorities

This paper describes the development of an optimization model for the management and operation of a large-scale, multireservoir water supply distribution system with preemptive priorities. The model considers multiobjectives and hedging rules. During periods of drought, when water supply is insufficient to meet the planned demand, appropriate rationing factors are applied to reduce water supply. In this paper, a water distribution system is formulated as a network and solved by the GAMS modeling system for mathematical programming and optimization. A user-friendly interface is developed to facilitate the manipulation of data and to generate graphs and tables for decision makers. The optimization model and its interface form a decision support system (DSS), which can be used to configure a water distribution system to facilitate capacity expansion and reliability studies. Several examples are presented to demonstrate the utility and versatility of the developed DSS under different supply and demand scenarios, including applications to one of the largest water supply systems in the world, the Sao Paulo Metropolitan Area Water Supply Distribution System in Brazil.

Recycling Krylov subspaces for efficient large-scale electrical impedance tomography

Electrical impedance tomography (EIT) captures images of internal features of a body. Electrodes are attached to the boundary of the body, low intensity alternating currents are applied, and the resulting electric potentials are measured. Then, based on the measurements, an estimation algorithm obtains the three-dimensional internal admittivity distribution that corresponds to the image. One of the main goals of medical EIT is to achieve high resolution and an accurate result at low computational cost. However, when the finite element method (FEM) is employed and the corresponding mesh is refined to increase resolution and accuracy, the computational cost increases substantially, especially in the estimation of absolute admittivity distributions. Therefore, we consider in this work a fast iterative solver for the forward problem, which was previously reported in the context of structural optimization. We propose several improvements to this solver to increase its performance in the EIT context. The solver is based on the recycling of approximate invariant subspaces, and it is applied to reduce the EIT computation time for a constant and high resolution finite element mesh. In addition, we consider a powerful preconditioner and provide a detailed pseudocode for the improved iterative solver. The numerical results show the effectiveness of our approach: the proposed algorithm is faster than the preconditioned conjugate gradient (CG) algorithm. The results also show that even on a standard PC without parallelization, a high mesh resolution (more than 150,000 degrees of freedom) can be used for image estimation at a relatively low computational cost. (C) 2010 Elsevier B.V. All rights reserved.

Large-scale ecological restoration of high-diversity tropical forests in SE Brazil

The complex interactions among endangered ecosystems, landowners` interests, and different models of land tenure and use, constitute an important series of challenges for those seeking to maintain and restore biodiversity and augment the flow of ecosystem services. Over the past 10 years, we have developed a data-based approach to address these challenges and to achieve medium and large-scale ecological restoration of riparian areas on private lands in the state of Sao Paulo, southeastern Brazil. Given varying motivations for ecological restoration, the location of riparian areas within landholdings, environmental zoning of different riparian areas, and best-practice restoration methods were developed for each situation. A total of 32 ongoing projects, covering 527,982 ha, were evaluated in large sugarcane farms and small mixed farms, and six different restoration techniques have been developed to help upscale the effort. Small mixed farms had higher portions of land requiring protection as riparian areas (13.3%), and lower forest cover of riparian areas (18.3%), than large sugarcane farms (10.0% and 36.9%, respectively for riparian areas and forest cover values). In both types of farms, forest fragments required some degree of restoration. Historical anthropogenic degradation has compromised forest ecosystem structure and functioning, despite their high-diversity of native tree and shrub species. Notably, land use patterns in riparian areas differed markedly. Large sugarcane farms had higher portions of riparian areas occupied by highly mechanized agriculture, abandoned fields, and anthropogenic wet fields created by siltation in water courses. In contrast, in small mixed crop farms, low or non-mechanized agriculture and pasturelands were predominant. Despite these differences, plantations of native tree species covering the entire area was by far the main restoration method needed both by large sugarcane farms (76.0%) and small mixed farms (92.4%), in view of the low resilience of target sites, reduced forest cover, and high fragmentation, all of which limit the potential for autogenic restoration. We propose that plantations should be carried out with a high-diversity of native species in order to create biologically viable restored forests, and to assist long-term biodiversity persistence at the landscape scale. Finally, we propose strategies to integrate the political, socio-economic and methodological aspects needed to upscale restoration efforts in tropical forest regions throughout Latin America and elsewhere. (C) 2010 Elsevier BA/. All rights reserved.