Minimum temperature forecast is particularly important for scheduling spring frost protection. The natural variability of minimum night temperature is strongly affected by the physical mechanisms ruling over energetic exchanges between soil and atmosphere. Under stable nocturnal thermal conditions (as in the case of frost nights), temperature differentiation patterns are created, strongly linked to the morphological features of individual areas. A high-resolution survey, like the one obtainable from thermal infrared (TIR) remote sensing, allows a temperature mapping, where differences among sites can be highlighted. In two frost-prone fruit-growing areas (Rotaliana Plain and lower Non Valley) in Trentino, northern Italy, airborne TIR surveys have been undertaken and the relevant thermal maps have been produced at a 10-m resolution. Calibration was carried out by contemporary ground measurements. Since emissivity remarkably changes depending on landcover features, an empirical calibration is proposed, carried out by ground temperature measurements, that enables to distinguish between any parcel, according to land use. The application to the real case led to temperature assessment errors of 0.95°C (as root of mean-squared errors, RMSE). Finally, in order to assess temperature at 2 m from its airborne TIR estimate at the ground level, statistical relationships have been investigated between values measured at the ground and at 2 m at some sample sites within the two target areas. Even if some outcomes are interesting and potentially useful, some inhomogeneity has resulted, partially hampering the general trial of temperature assessment at the conventional height of 2 m. Further investigation might lead to a fine-tuning of the procedures of temperature measurement and assessment. The methodology is proposed for a minimum temperature prediction at a high territorial resolution: a ground temperature pattern could be applied to reproduce the thermal anomaly referred to a site where an independently calibrated forecast is already known.
La previsione delle temperature minime è particolarmente importante nella programmazione delle azioni di difesa dalle gelate primaverili. La variabilità naturale della temperatura minima notturna è fortemente determinata dai meccanismi fisici che regolano gli scambi energetici tra suolo ed atmosfera. In condizioni di stabilità termica notturna (caso di interesse per le gelate) si creano pattern di differenziazione termica che sono fortemente legati alle caratteristiche morfologiche delle singole aree. Un rilievo ad elevato dettaglio territoriale, come quello ricavabile da telerilevamento nell'infrarosso termico (TIR), consente di mappare le temperature evidenziando, all'interno di un'area limitata, differenze tra sito e sito. In due aree frutticole trentine soggette ad episodi di gelo (Piana Rotaliana e Bassa Val di Non, Trentino) sono stati eseguiti rilievi aerei nella banda del TIR e sono state ricavate le relative mappe termiche con risoluzione 10 m. La calibrazione è stata eseguita mediante contemporanee misure a terra di temperatura. Poiché l'emissività varia anche considerevolmente a seconda dell'uso del suolo, si propone una calibrazione empirica, condotta con dati di temperatura misurati al suolo, che consenta di distinguere tra parcella e parcella a seconda della sua destinazione d'uso. L'applicazione al caso reale ha portato a stime della temperatura con un errore di 0.95°C (come radice dei quadrati dei residui, RMSE). Per ricavare, infine, la temperatura a 2 m a partire da quella stimata al suolo con telerilievo aereo di TIR, si sono indagate le relazioni statistiche tra i valori misurati alle due diverse altezze in alcuni siti campione nelle stesse due aree di indagine. Pur ricavando informazioni di interesse per il loro potenziale utilizzo, ne è emersa una certa disomogeneità, che in parte ostacola la generalizzazione del procedimento di stima di temperatura all'altezza convenzionale di 2 m. Ulteriori approfondimenti potranno portare ad un affinamento delle procedure di misura e stima della temperatura. La metodologia proposta risulta utile per una previsione di temperatura minima ad elevato dettaglio territoriale: si tratterebbe di applicare una "maschera" che riproduca il modello di anomalia termica riferita ad un sito per il quale sia nota una previsione, indipendentemente calibrata.
Thermal infrared remote sensing for high-resolution minimum temperature mapping
Gioli B
2008
Abstract
Minimum temperature forecast is particularly important for scheduling spring frost protection. The natural variability of minimum night temperature is strongly affected by the physical mechanisms ruling over energetic exchanges between soil and atmosphere. Under stable nocturnal thermal conditions (as in the case of frost nights), temperature differentiation patterns are created, strongly linked to the morphological features of individual areas. A high-resolution survey, like the one obtainable from thermal infrared (TIR) remote sensing, allows a temperature mapping, where differences among sites can be highlighted. In two frost-prone fruit-growing areas (Rotaliana Plain and lower Non Valley) in Trentino, northern Italy, airborne TIR surveys have been undertaken and the relevant thermal maps have been produced at a 10-m resolution. Calibration was carried out by contemporary ground measurements. Since emissivity remarkably changes depending on landcover features, an empirical calibration is proposed, carried out by ground temperature measurements, that enables to distinguish between any parcel, according to land use. The application to the real case led to temperature assessment errors of 0.95°C (as root of mean-squared errors, RMSE). Finally, in order to assess temperature at 2 m from its airborne TIR estimate at the ground level, statistical relationships have been investigated between values measured at the ground and at 2 m at some sample sites within the two target areas. Even if some outcomes are interesting and potentially useful, some inhomogeneity has resulted, partially hampering the general trial of temperature assessment at the conventional height of 2 m. Further investigation might lead to a fine-tuning of the procedures of temperature measurement and assessment. The methodology is proposed for a minimum temperature prediction at a high territorial resolution: a ground temperature pattern could be applied to reproduce the thermal anomaly referred to a site where an independently calibrated forecast is already known.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
