ECOWHEATALY: calibration dataset for an agent-based model of Italian durum wheat farms based on RICA and ISTAT census

Description This dataset was developed as part of the Ecowheataly Research Project: "Evaluation of policies for enhancing sustainable wheat production in Italy" (202288L9YN). The European Union funds the project - Next Generation EU, through a call issued by the Minister of University and Research for projects of significant national interest (PRIN 2022). One of the goals of Ecowheataly is to represent the Italian wheat production system at the micro level and to scale it up to the sectoral level while preserving empirically observed heterogeneity. The core idea is to start from a microeconomic description of an individual farm - grounded in profit maximization under agronomic and economic constraints - and then construct an ensemble of heterogeneous farms such that the simulated population reproduces key structural and technological features of Italian wheat production (e.g., dispersion in yields, input intensities, and farm size patterns). Methodologically, the framework follows the Agent-Based computational economics paradigm, which studies aggregate outcomes as emergent properties of decentralized decision-making by heterogeneous agents interacting within a defined institutional environment. In this framework, parameter values that enter the microeconomic farm model are estimated using both Italian farm accounting microdata (RICA, Agricultural Accounting Information Network) and structural information from the 2020 Agricultural census (ISTAT, Italian National Institute of Statistics). The RICA survey is the harmonized European source for farm management and economic variables, coordinated in Italy by the Council for Research and Analysis of Agricultural Economics (CREA). On the other hand, the ISTAT agricultural census provides a comprehensive picture of the structure of Italian agricultural holdings at the national and sub-national levels. Elaborations built from such microdata form the Ecowheataly dataset. The Ecowheataly dataset is composed of 2 blocks: Block 1 derived from RICA database Block 2 derived from ISTAT - 7th General Census of Agriculture Block 1 (estimation_all.csv) RICA is a comprehensive Italian farm database managed by the Italian Council for Research and Analysis of Agricultural Economics (CREA). RICA is the Italian counterpart to the Farm Accountancy Data Network (FADN), an annual sample survey established by the European Commission to inform the Common Agricultural Policy. The RICA system annually collects survey data on approximately 12000 farms, of which more than 3000 are identified as wheat producers. It provides quantitative insights into agricultural practices at the individual farm level across all regions of Italy. Within the Ecowheataly project, we model a situation in which a product (wheat) is produced using three inputs: Nitrogen, Herbicide, and Insecticide. In particular, we adopt the economic principle that choosing an input combination is formulated as a cost- minimization problem, that is, the farmer selects the input combination that minimizes total cost. Methodological details are described in the Technical Report. The first block of the Ecowheataly dataset is derived from the roughly 30000 records of RICA for the period 2008-2022 and is structured as follows: 432 records and 20 variables (rows x columns) 3 spatial levels (1 NUT 1; 20 NUT 2; 86 NUT 3) 4 elevation levels (any altitude; hill =' Collina`; mountain =' Montagna`; plain =' Pianura`) The remaining variables are described as follows: s1, s2, and s3 indicate the share of potential yield lost due to stress associated with the use of nitrogen (need for fertilizers), herbicide (biotic stress caused by weeds), and insecticide (biotic stress caused by insect pests) lambda1, lambda2, and lambda3 represent estimates of the parameters governing the yield recovery function under the use of the input max _ yield _ s1 is the estimate of the maximum yield attainable by using the optimal level of nitrogen risk measures the uncertainty of input effectiveness and is estimated using the standard deviation of the wheat yields recorded at the specific spatial and elevation levels N _ cases _ XX reports the number of records used for the estimation phase, unless otherwise indicated in the column Flag _ XX Flag _ XX indicates the spatial level of the estimation process (if data for a specific province and elevation are not useful for obtaining the estimates, then the higher NUT level is used). Block 2 (Cens-gen-Distribution-per-Province-parameters.csv) The second block of the Ecowheataly dataset is derived from two census datasets: Census of land localization ("Censimento localizzazione terreni") General farm census dataset ("Censimento dati generali"). The microdata used to obtain the Ecowheataly dataset are from the 7th Italian Agricultural Census (2020), which is managed by the National Institute of Statistics-ISTAT. Access to census microdata was granted through the ISTAT ADELE Laboratory within the ECOWHEATALY project framework From the land localization dataset, 195735 durum wheat-producing farms were identified and aggregated by province/region and altimetry class. From the general census dataset, 136041 durum wheat-producing farms were used for demographic and structural aggregation. Moreover, altimetry classes were harmonized into three categories: a) mountain, b) hill, and c) plain. While methodological details are described in the Technical Report, the Table of data derived from microdata elaborations is named Cens-gen-Distribution-per-Province-parameters.csv and is structured as follows: 195 records and 21 variables (rows x columns) 1 spatial level (99 NUT 3) 3 elevation levels (hill =' Collina`; mountain =' Montagna`; plain =' Pianura`) Note that durum wheat is cropped in many places but not everywhere; that is why not all the provinces or elevations are represented in the dataset. The remaining variables are described as follows: DistrName_SAU, together with parXX_SAU, defines the estimated statistical distribution of microdata and associated parameters for the whole cropped area of the farm DistrName_CER2, together with parXX_CER2, represents the estimated statistical distribution and associated parameters for the area devoted to durum wheat cropping DistrName_AGE, together with parXX_AGE, represents the estimated statistical distribution and associated parameters for the farm manager's age. The data used in this work are from the Istat source and relate to the 7th General Census of Agriculture. The elaborations were conducted at the Laboratory for the Analysis of Istat Elementary Data and in compliance with the legislation on the protection of statistical confidentiality and the protection of personal data. The results and opinions expressed are the sole responsibility of the author and do not constitute official statistics. It should be noted that the analyses were conducted without using the weights reported to the universe.