ArcheoClima Database: A Relational Dataset of Historical Climate Events in Southern Italy (6th–19th Century CE)

This dataset is the companion database to the article "Carletti, E., Ragno, R. 2026. Cambiamenti climatici nel Mezzogiorno adriatico tra I e II millennio d.C.: fonti scritte e modelli paleoclimatici. In: ArcheoClima. Edipuglia, Bari". It contains 526 documented historical climate events for central-southern Italy (principally Puglia, with additional coverage of Basilicata, Molise, Campania, Abruzzo, and southern Lazio), spanning approximately 600–1900 CE. The analytical focus of the associated study is the period 1000–1800 CE, when the density of events is higher. Events were extracted from a wide range of documentary sources — monastic annals, chronicles, notarial acts, diplomatic registers, Angevin administrative records, and early modern scientific treatises — and classified according to event type, intensity, spatial extent, and chronological precision. The dataset was designed to be compatible with Pfister-scale coding for integration with CHELSA-TraCE21k palaeoclimatic simulations via Bayesian modelling. Analysis scripts are available at: github.com/robertoragno/archeoclima-reconstructions. Files File Description Dataset_EN.csv Main events table (526 rows, one per documented climate event). There is also an italian version available in the same folder, named Dataset_IT.csv Tbl_Event_Types.xlsx Dictionary of the 18 event type codes Tbl_Temporal_Precision.xlsx Dictionary of the 5 chronological precision levels Tbl_Sources.xlsx Documentary and bibliographic sources (one row per source attestation; multiple rows per event are possible) Tbl_Source_Type_Dictionary.xlsx Dictionary of the 10 source type codes All files are UTF-8 encoded XSLX, except Dataset_EN.csv which is comma delimited for simplicity in the analysis. Missing values are represented as empty cells. Database Structure The database is organised as five relational tables. Dataset_EN.csv is the central table; the remaining four are lookup dictionaries linked to it by numeric codes. Tbl_Sources.csv has a one-to-many relationship with the main table: a single event may be attested by multiple independent sources, each stored as a separate row. Dataset.csv — Main Events Table Field Description ID_Event Unique progressive identifier for each recorded event Event_Type Event type code (1–16); links to Tbl_Event_Types.csv Event_Description Free-text narrative description of the event Event_Strength Ordinal intensity: 1 = Uncertain, 2 = Moderate, 3 = Strong, 4 = Extreme Geographic_Impact Spatial extent: 1 = Local, 2 = Sub-regional, 3 = Regional, 4 = Macro-territorial Start_Date Start year (yyyy) End_Date End year (yyyy); equals start year for point events Precise_Start_Date Precise start date (dd/mm/yyyy), where the source provides this resolution Precise_End_Date Precise end date (dd/mm/yyyy), where available Season Season: Spring, Summer, Autumn, Winter Chronological_Precision Chronological precision level (1–5); links to Tbl_Temporal_Precision.csv Lat Latitude (decimal degrees), where point georeferencing is possible Lon Longitude (decimal degrees) Modern_Location_Name Modern place name Ancient_Location_Name Historical place name as attested in the source Spatial_Precision TRUE if geolocation is point-precise; FALSE if only regional attribution is possible Extra_Notes Additional methodological notes or cross-references Tbl_Sources.xlsx — Sources Table One row per source attestation. Linked to the main table via ID_Event = ID_Event. A single event may have multiple rows here. Field Description Source_ID Unique progressive identifier for each source record Event_ID Foreign key linking to ID_Event in the main table Ancient_Source Reference to the primary historical source (e.g. Paolo Diacono, Historia Langobardorum, II, 25) Ancient_Author Author of the source Modern_Reference Full bibliographic reference for the modern critical edition consulted Source_reliability Critical reliability of the source regarding the reported event (TRUE/FALSE) Source_Type Source type code (1–10); links to Tbl_Source_Type_Dictionary.csv Source_Text Verbatim transcription of the relevant passage from the primary source Lookup Tables File Key field Contents Tbl_Event_Types.xlsx Event_Type_ID Maps codes 1–16 to event type descriptions (e.g. drought, flood, hailstorm, extreme cold, locust invasion, famine). Types 1, 3, 4, 5, 8, and 9 are the strictly climatic events used for Pfister-scale coding in the associated analysis. Tbl_Temporal_Precision.xlsx Precision level ID Maps levels 1–5, from uncertain (decade-level margin) to precise to the day. 99.2% of events are dateable to at least the year (levels 3–5); 59% are dateable to the day. Tbl_Source_Type_Dictionary.xlsx Source_ID Maps codes 1–10 to source type categories (e.g. monastic annals, chronicles, notarial acts, Angevin registers, scientific treatises). Summary of Dataset Contents Temporal distribution The 18th century is the most densely documented period (237 events), largely due to the systematic meteorological observations of Giuseppe Maria Giovene and the greater volume of modern administrative records. The 7th–10th centuries are sparsely represented, partly because of differential source survival and partly due to the destruction of the Angevin archive during World War II. The 16th and 17th centuries are also underrepresented, as a comprehensive survey of available sources for that period has not yet been completed. Geographical distribution Coverage is heaviest along the Adriatic coast of Puglia and in areas historically tied to major documentary centres (Bari, Foggia, Brindisi, Molfetta, Lecce). The Montecassino monastic archive contributes significant coverage for southern Lazio and the Campania–Molise border area. Inland areas such as the Sub-Apennine Dauno and the Lucanian Apennines are more sparsely documented. A portion of events — those attributed generically to "Puglia" or "the Kingdom" — cannot be assigned a precise point location and carry Spatial_Precision = FALSE. Event types Strictly climatic events (Types 1, 3, 4, 5, 8, 9) account for approximately 54% of all records (N = 285). Among these, events indicating excess precipitation or cold conditions are more frequent than those indicating heat or drought, which partly reflects the selective nature of historical documentation — extreme cold and flooding tended to produce more durable written traces in this region. Famine (Type 7) is the most frequently recorded event type overall (N = 98), retained in the database as a potentially climate-correlated indicator but excluded from the Pfister-scale analysis given its complex social causation. Chronological precision 99.2% of events are dateable to at least the year (precision levels 3–5). 59% are dateable to the day. Despite this high individual precision, the uneven distribution of events across centuries means that centennial aggregation is the appropriate resolution for quantitative analysis. Joining the Tables In R: library(readxl) events <- read.csv("Dataset_EN.csv") ev_types <- read_excel("Tbl_Event_Types.xlsx") sources <- read_excel("Tbl_Sources.xlsx") # Add event type labels events <- merge(events, ev_types, by.x = "Event_Type", by.y = "Event_Type_ID") # Retrieve all sources for each event (one-to-many) events_with_sources <- merge(events, sources, by.x = "ID_Event", by.y = "ID_Event") # Filter to climatic events dateable to the year or better climatic <- subset(events, Event_Type %in% c(1, 3, 4, 5, 8, 9) & Chronological_Precision >= 3) Funding Progetto BAC "ArcheoCLIMA. Archeologia dei Cambiamenti Climatici nel Mezzogiorno Antico" — CUP HS3C22000860006, funded within the project "CHANGES – Cultural Heritage Active Innovation for Sustainable Society" — Codice PE00000020 — CUP H53C22000860006 — Spoke 1 "Historical Landscapes, Traditions and Cultural Identities" — PNRR Missione 4 "Istruzione e ricerca", Componente 2 "Dalla ricerca all'impresa", Investimento 1.3 "Partenariati estesi a Università, centri di ricerca, imprese e finanziamento progetti di ricerca" (Unione Europea – Next Generation EU). PI: Università del Salento. Partner: Università di Foggia; Webgenesys spa.