Peach leaf curl, caused by the fungus Taphrina deformans, is a worldwide disease of peach and nectarine. Symptoms include leaf thickening, curling, discoloration and can finally lead to severe defoliation and reduced productivity. Current management relies on pre- ventive treatments, often supported by empirical models based on weather thresholds and simplified phenological windows, with limited transferability across years and differ- ent climatic areas. This thesis develops a mechanistic, weather-driven epidemiological model that integrates host phenology, infection pressure, and meteorological conditions within an SEIR-type framework applied to leaves. The model is forced by time series of temperature and leaf wetness and simulates canopy fractions in four compartments (S susceptible, E exposed asymptomatic, I symptomatic, R resistant). The output is the seasonal dynamics of symptom severity, expressed as the percentage of symptomatic leaves over adult leaves, consistent with the metric used in field assessments. Parame- ters are estimated sequentially by combining literature evidence and observed data: leaf emergence, exposure favorability, and latency components are parameterized using bi- ological constraints and experimental data, while two key epidemiological parameters are calibrated on severity data from the INRAE orchard (Saint-Marcel-lès-Valence, 2025) through minimization of the RMSE between observed and simulated series. Model perfor- mance is discussed through a pattern-oriented evaluation (timing and shape of dynamics, weather–phenology–disease relationships) and supported by error metrics used for calibra- tion purposes. Overall, simulations reproduce robust features of canopy dynamics and the growth and decline of severity, while indicating possible improvements in the post-peak phase related to processes of removal or closure of susceptibility.
La bolla del pesco, causata dal fungo Taphrina deformans, è una malattia di pesco e nettarino diffusa a livello mondiale. I sintomi sono ispessimento, increspamento e decolo- razione delle foglie; la patologia può determinare una forte defogliazione e una conseguente riduzione della produttività. La gestione attuale si basa su trattamenti preventivi, spesso supportati da modelli empirici fondati su soglie meteo e finestre fenologiche semplificate, con limitata trasferibilità tra anni e aree climatiche differenti. Questa tesi sviluppa un modello epidemiologico meccanicistico, climate-driven, che integra fenologia dell’ospite, pressione d’infezione e condizioni meteorologiche in un quadro tipo SEIR applicato alle foglie. Il modello prende in ingresso serie temporali di temperatura dell’aria e bagnatura fogliare, simula le frazioni di chioma in quattro compartimenti (S suscettibili, E espo- ste asintomatiche, I sintomatiche, R resistenti). L’output è la dinamica stagionale della severità dei sintomi, espressa come percentuale di foglie sintomatiche sulle foglie adulte, coerente con la metrica utilizzata nei rilievi di campo. I parametri sono stimati in modo sequenziale combinando evidenze di letteratura e dati osservati: le componenti di emergen- za fogliare, favorabilità d’esposizione e latenza sono parametrizzate con vincoli biologici e dati sperimentali, mentre due parametri epidemiologici chiave sono calibrati sui dati di severità del frutteto INRAE (Saint-Marcel-lès-Valence, 2025) tramite minimizzazione dell’RMSE tra serie osservate e simulate. La performance del modello è discussa tra- mite una valutazione orientata ai pattern (tempistica e forma delle dinamiche, relazioni meteo-fenologia-malattia) e supportata da metriche di errore usate a fini di calibrazio- ne. Nel complesso, le simulazioni riproducono caratteristiche robuste delle dinamiche di chioma e l’andamento di crescita e declino della severità, indicando anche possibili miglio- ramenti nella fase post-picco legati ai processi di rimozione/chiusura della suscettibilità.
A weather-driven SEIR-like model for peach leaf curl: application to a southern France orchard
Cassella, Daniela
2025/2026
Abstract
Peach leaf curl, caused by the fungus Taphrina deformans, is a worldwide disease of peach and nectarine. Symptoms include leaf thickening, curling, discoloration and can finally lead to severe defoliation and reduced productivity. Current management relies on pre- ventive treatments, often supported by empirical models based on weather thresholds and simplified phenological windows, with limited transferability across years and differ- ent climatic areas. This thesis develops a mechanistic, weather-driven epidemiological model that integrates host phenology, infection pressure, and meteorological conditions within an SEIR-type framework applied to leaves. The model is forced by time series of temperature and leaf wetness and simulates canopy fractions in four compartments (S susceptible, E exposed asymptomatic, I symptomatic, R resistant). The output is the seasonal dynamics of symptom severity, expressed as the percentage of symptomatic leaves over adult leaves, consistent with the metric used in field assessments. Parame- ters are estimated sequentially by combining literature evidence and observed data: leaf emergence, exposure favorability, and latency components are parameterized using bi- ological constraints and experimental data, while two key epidemiological parameters are calibrated on severity data from the INRAE orchard (Saint-Marcel-lès-Valence, 2025) through minimization of the RMSE between observed and simulated series. Model perfor- mance is discussed through a pattern-oriented evaluation (timing and shape of dynamics, weather–phenology–disease relationships) and supported by error metrics used for calibra- tion purposes. Overall, simulations reproduce robust features of canopy dynamics and the growth and decline of severity, while indicating possible improvements in the post-peak phase related to processes of removal or closure of susceptibility.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_03_Cassella_Tesi.pdf
accessibile in internet per tutti a partire dal 02/03/2029
Descrizione: Testo della tesi
Dimensione
5.57 MB
Formato
Adobe PDF
|
5.57 MB | Adobe PDF | Visualizza/Apri |
|
2026_03_Cassella_Executive_Summary.pdf
accessibile in internet per tutti a partire dal 02/03/2029
Descrizione: Executive summary della tesi
Dimensione
1.13 MB
Formato
Adobe PDF
|
1.13 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/252686