Roberto Benedetti

On the use of NDVI profiles as a tool for agricultural statistics: The case study of wheat yield estimate and forecast in Emilia Romagna

Abstract The use of satellite-derived vegetation indices for crop monitoring and yield estimate and forecast is of key importance for those organizations in charge to monitor the agrarian season. This study intends to investigate the potential use of AVHRR / NDVI data for wheat monitoring in Italy. The time frame chosen is the 4-year span between 1986 and 1989, and the study region considered is Emilia Romagna. Two scales of study have been used: microscale and mesoscale. The first scale corresponds to the limits of NOAA satellite spatial resolution, and has been used in the study of the vegetation index on restricted test sites, which, nevertheless, revealed a large number of data, including ground coverage. A wider scale has been considered to extend the results obtained in the microlevel analysis to the lowland section of the Emilia Romagna region. Good correlations were found with ground simulated and collected crop parameters. In particular, NDVI has been found to be highly representative of plant photosynthetic capacity and efficiency. Using these results, a simple linear regression model has been derived for wheat yield estimate and forecast based on NDVI integration during the wheat grain filling period. The results obtained, compared with official data, show their usefulness for a cheap and real-time crop monitoring.

Vegetation classification in the middle Mediterranean area by satellite data

Abstract The Normalized Difference Vegetation Index time series derived from NOAA satellite data relevant to the years 1986–1989 have been used to evaluate their usefulness for vegetation mapping purposes. Given the remote sensing information about all the available dates in a period, it is possible to define areas of homogeneous time NDVI profile through unsupervised classification procedures. In the discrimination process it is assumed that dynamic features of the vegetation with an evolution period shorter than one month can be neglected. Three different approaches are used: an unsupervised classification on Principal Components and on 6 and 12 months data. Maximum value composites of the NDVI profiles (in LAC format) in each area for each month of the four-year time frame have proved to be very useful to stress the various typologies of vegetation development and condition. The output vegetation map is characterized by regions of a similar vegetation composition and timing among the different dominant...

Agricultural survey methods

List of Contributors. Introduction. 1 The present state of agricultural statistics in developed countries: situation and challenges. 1.1 Introduction. 1.2 Current state and political and methodological context. 1.3 Governance and horizontal issues. 1.4 Development in the demand for agricultural statistics. 1.5 Conclusions. Acknowledgements. Reference. Part I Census, Frames, Registers and Administrative Data. 2 Using administrative registers for agricultural statistics. 2.1 Introduction. 2.2 Registers, register systems and methodological issues. 2.3 Using registers for agricultural statistics. 2.4 Creating a farm register: the population. 2.5 Creating a farm register: the statistical units. 2.6 Creating a farm register: the variables. 2.7 Conclusions. References. 3 Alternative sampling frames and administrative data. What is the best data source for agricultural statistics? 3.1 Introduction. 3.2 Administrative data. 3.3 Administrative data versus sample surveys. 3.4 Direct tabulation of administrative data. 3.5 Errors in administrative registers. 3.6 Errors in administrative data. 3.7 Alternatives to direct tabulation. 3.8 Calibration and small-area estimators. 3.9 Combined use of different frames. 3.10 Area frames. 3.11 Conclusions. Acknowledgements. References. 4 Statistical aspects of a census. 4.1 Introduction. 4.2 Frame. 4.3 Sampling. 4.4 Non-sampling error. 4.5 Post-collection processing. 4.6 Weighting. 4.7 Modelling. 4.8 Disclosure avoidance. 4.9 Dissemination. 4.10 Conclusions. References. 5 Using administrative data for census coverage. 5.1 Introduction. 5.2 Statistics Canada s agriculture statistics programme. 5.3 1996 Census. 5.4 Strategy to add farms to the farm register. 5.5 2001 Census. 5.6 2006 Census. 5.7 Towards the 2011 Census. 5.8 Conclusions. Acknowledgements. References. Part II Sample Design, Weighting and Estimation. 6 Area sampling for small-scale economic units. 6.1 Introduction. 6.2 Similarities and differences from household survey design. 6.3 Description of the basic design. 6.4 Evaluation criterion: the effect of weights on sampling precision. 6.5 Constructing and using strata of concentration . 6.6 Numerical illustrations and more flexible models. 6.7 Conclusions. Acknowledgements. References. 7 On the use of auxiliary variables in agricultural survey design. 7.1 Introduction. 7.2 Stratification. 7.3 Probability proportional to size sampling. 7.4 Balanced sampling. 7.5 Calibration weighting. 7.6 Combining ex ante and ex post auxiliary information: a simulated approach. 7.7 Conclusions. References. 8 Estimation with inadequate frames. 8.1 Introduction. 8.2 Estimation procedure. References. 9 Small-area estimation with applications to agriculture. 9.1 Introduction. 9.2 Design issues. 9.3 Synthetic and composite estimates. 9.4 Area-level models. 9.5 Unit-level models. 9.6 Conclusions. References. Part III GIS and Remote Sensing. 10 The European land use and cover area-frame statistical survey. 10.1 Introduction. 10.2 Integrating agricultural and environmental information with LUCAS. 10.3 LUCAS 2001 2003: Target region, sample design and results. 10.4 The transect survey in LUCAS 2001 2003. 10.5 LUCAS 2006: a two-phase sampling plan of unclustered points. 10.6 Stratified systematic sampling with a common pattern of replicates. 10.7 Ground work and check survey. 10.8 Variance estimation and some results in LUCAS 2006. 10.9 Relative efficiency of the LUCAS 2006 sampling plan. 10.10 Expected accuracy of area estimates with the LUCAS 2006 scheme. 10.11 Non-sampling errors in LUCAS 2006. 10.12 Conclusions. Acknowledgements. References. 11 Area frame design for agricultural surveys. 11.1 Introduction. 11.2 Pre-construction analysis. 11.3 Land-use stratification. 11.4 Sub-stratification. 11.5 Replicated sampling. 11.6 Sample allocation. 11.7 Selection probabilities. 11.8 Sample selection. 11.9 Sample rotation. 11.10 Sample estimation. 11.11 Conclusions. 12 Accuracy, objectivity and efficiency of remote sensing for agricultural statistics. 12.1 Introduction. 12.2 Satellites and sensors. 12.3 Accuracy, objectivity and cost-efficiency. 12.4 Main approaches to using EO for crop area estimation. 12.5 Bias and subjectivity in pixel counting. 12.6 Simple correction of bias with a confusion matrix. 12.7 Calibration and regression estimators. 12.8 Examples of crop area estimation with remote sensing in large regions. 12.9 The GEOSS best practices document on EO for crop area estimation. 12.10 Sub-pixel analysis. 12.11 Accuracy assessment of classified images and land cover maps. 12.12 General data and methods for yield estimation. 12.13 Forecasting yields. 12.14 Satellite images and vegetation indices for yield monitoring. 12.15 Examples of crop yield estimation/forecasting with remote sensing. References. 13 Estimation of land cover parameters when some covariates are missing. 13.1 Introduction. 13.2 The AGRIT survey. 13.3 Imputation of the missing auxiliary variables. 13.4 Analysis of the 2006 AGRIT data. 13.5 Conclusions. References. Part IV Data Editing and Quality Assurance. 14 A generalized edit and analysis system for agricultural data. 14.1 Introduction. 14.2 System development. 14.3 Analysis. 14.4 Development status. 14.5 Conclusions. References. 15 Statistical data editing for agricultural surveys. 15.1 Introduction. 15.2 Edit rules. 15.3 The role of automatic editing in the editing process. 15.4 Selective editing. 15.5 An overview of automatic editing. 15.6 Automatic editing of systematic errors. 15.7 The Fellegi Holt paradigm. 15.8 Algorithms for automatic localization of random errors. 15.9 Conclusions. References. 16 Quality in agricultural statistics. 16.1 Introduction. 16.2 Changing concepts of quality. 16.3 Assuring quality. 16.4 Conclusions. References. 17 Statistics Canada s Quality Assurance Framework applied to agricultural statistics. 17.1 Introduction. 17.2 Evolution of agriculture industry structure and user needs. 17.3 Agriculture statistics: a centralized approach. 17.4 Quality Assurance Framework. 17.5 Managing quality. 17.6 Quality management assessment. 17.7 Conclusions. Acknowledgements. References. Part V Data Dissemination and Survey Data Analysis. 18 The data warehouse: a modern system for managing data. 18.1 Introduction. 18.2 The data situation in the NASS. 18.3 What is a data warehouse? 18.4 How does it work? 18.5 What we learned. 18.6 What is in store for the future? 18.7 Conclusions. 19 Data access and dissemination: some experiments during the First National Agricultural Census in China. 19.1 Introduction. 19.2 Data access and dissemination. 19.3 General characteristics of SDA. 19.4 A sample session using SDA. 19.5 Conclusions. References. 20 Analysis of economic data collected in farm surveys. 20.1 Introduction. 20.2 Requirements of sample surveys for economic analysis. 20.3 Typical contents of a farm economic survey. 20.4 Issues in statistical analysis of farm survey data. 20.5 Issues in economic modelling using farm survey data. 20.6 Case studies. References. 21 Measuring household resilience to food insecurity: application to Palestinian households. 21.1 Introduction. 21.2 The concept of resilience and its relation to household food security. 21.3 From concept to measurement. 21.4 Empirical strategy. 21.5 Testing resilience measurement. 21.6 Conclusions. References. 22 Spatial prediction of agricultural crop yield. 22.1 Introduction. 22.2 The proposed approach. 22.3 Case study: the province of Foggia. 22.4 Conclusions. References. Author Index. Subject Index.

A regression tree algorithm for the identification of convergence clubs

The concept of convergence clubs is analyzed and compared with classical methods for the study of economic @b-convergence, which often consider the entire data set as one sample. A technique for the identification of convergence clubs is proposed. The algorithm is based on a modified version of the usual regression trees procedure. The objective function of the method is represented by the difference among the parameters of the model under investigation. Different strategies are adopted in the definition of the model used in the objective function of the algorithm. The first is the classical non-spatial @b-convergence model. The others are modified @b-convergence models which take into account the dependence showed by spatially distributed data. The proposed procedure identifies situation of local stationarity in the economic growth of the different regions: a group of regions is divided into two sub-groups if the parameter estimates are significantly different among them. The algorithm is applied to 191 European regions for the period 1980-2002. Given the adaptability of the algorithm, its implementation provides a flexible tool for the use of any regression model in the analysis of non-stationary spatial data.

A Tree-Based Approach to Forming Strata in Multipurpose Business Surveys

The design of a stratified sample from a finite population deals with two main issues: the definition of a rule to partition the population, and the allocation of sampling units in the selected strata. This article examines a tree-based strategy which plans to solve jointly these issues when the survey is multipurpose and multivariate information, quantitative or qualitative, is available. Strata are formed through a scissorial algorithm that selects finer and finer partitions by minimizing, at each step, the sample allocation required to achieve the precision levels set for each surveyed variable. In this way, large numbers of constraints can be satisfied without drastically increasing the sample size, and also without discarding variables selected for stratification or diminishing the number of their class intervals. Furthermore, the algorithm tends to not define empty or almost empty strata, so avoiding the need for ex post strata aggregations. The procedure was applied to redesign the Italian Farm Structure Survey. The results indicate that the gain in efficiency held using our strategy is nontrivial. For a given sample size, this procedure achieves the required precision by exploiting a number of strata which is usually a very small fraction of the number of strata available when combining all possible classes from any of the covariates.

A two-step approach to account for unobserved spatial heterogeneity

ABSTRACT A two-step approach to account for unobserved spatial heterogeneity. Spatial Economic Analysis. Empirical analysis in economics often faces the difficulty that the data are correlated and heterogeneous in some unknown form. Spatial econometric models have been widely used to account for dependence structures, but the problem of directly dealing with unobserved spatial heterogeneity has been largely unexplored. The problem can be serious particularly if we have no prior information justified by economic theory. In this paper we propose a two-step procedure to identify endogenously spatial regimes in the first step and to account for spatial dependence in the second step. This procedure is applied to hedonic house price analysis.

Approximate maximum likelihood estimation of the Bingham distribution

Maximum likelihood estimation of the Bingham distribution is difficult because the density function contains a normalization constant that cannot be computed in closed form. Given the availability of sufficient statistics, Approximate Maximum Likelihood Estimation (AMLE) is an appealing method that allows one to bypass the evaluation of the likelihood function. The impact of the input parameters of the AMLE algorithm is investigated and some methods for choosing their numerical values are suggested. Moreover, AMLE is compared to the standard approach which numerically maximizes the (approximate) likelihood obtained with the normalization constant estimated via the Holonomic Gradient Method (HGM). For the Bingham distribution on the sphere, simulation experiments and real-data applications produce similar outcomes for both methods. On the other hand, AMLE outperforms HGM when the dimension increases.

A Bayesian Approach to Parameter Estimation in the Presence of Spatial Missing Data

Abstract The missing data problem has been widely addressed in the literature. The traditional methods for handling missing data may be not suited to spatial data, which can exhibit distinctive structures of dependence and/or heterogeneity. As a possible solution to the spatial missing data problem, this paper proposes an approach that combines the Bayesian Interpolation method [Benedetti, R. & Palma, D. (1994) Markov random field-based image subsampling method, Journal of Applied Statistics, 21(5), 495–509] with a multiple imputation procedure. The method is developed in a univariate and a multivariate framework, and its performance is evaluated through an empirical illustration based on data related to labour productivity in European regions.

Local stationarity in small area estimation models

Small area estimators are often based on linear mixed models under the assumption that relationships among variables are stationary across the area of interest (Fay–Herriot models). This hypothesis is patently violated when the population is divided into heterogeneous latent subgroups. In this paper we propose a local Fay–Herriot model assisted by a Simulated Annealing algorithm to identify the latent subgroups of small areas. The value minimized through the Simulated Annealing algorithm is the sum of the estimated mean squared error (MSE) of the small area estimates. The technique is employed for small area estimates of erosion on agricultural land within the Rathbun Lake Watershed (IA, USA). The results are promising and show that introducing local stationarity in a small area model may lead to useful improvements in the performance of the estimators.

Modelling spatial regimes in farms technologies

We exploit the information derived from geographical coordinates to endogenously identify spatial regimes in technologies that are the result of a variety of complex, dynamic interactions among site-specific environmental variables and farmer decision making about technology, which are often not observed at the farm level. Controlling for unobserved heterogeneity is a fundamental challenge in empirical research, as failing to do so can produce model misspecification and preclude causal inference. In this article, we adopt a two-step procedure to deal with unobserved spatial heterogeneity, while accounting for spatial dependence in a cross-sectional setting. The first step of the procedure takes explicitly unobserved spatial heterogeneity into account to endogenously identify subsets of farms that follow a similar local production econometric model, i.e. spatial production regimes. The second step consists in the specification of a spatial autoregressive model with autoregressive disturbances and spatial regimes. The method is applied to two regional samples of olive growing farms in Italy. The main finding is that the identification of spatial regimes can help drawing a more detailed picture of the production environment and provide more accurate information to guide extension services and policy makers.