In 2014, the pine tortoise scale (Toumeyella parvicornis Cockerell) was detected for the first time in Italy, causing severe infestations in Pinus pinea L. forests, particularly in Naples and Rome. This research aims to: identify the damage caused by T. parvicornis in selected green areas of the city of Rome; test the effectiveness of the evaluation sheet developed for detecting infestations; assess the efficiency of images from the PlanetScope constellation in monitoring the impact of T. parvicornis on stone pines (P. pinea); and analyze the advantages and challenges of integrating field surveys and remote sensing in detecting, monitoring, and managing insect attacks on urban trees. The studies published so far on the attack of the pine tortoise scale on P. pinea have been conducted mainly in some areas of Rome and Naples. Among these, those based on remote sensing predominantly focus on suburban areas or extensive pine forests, while others monitor infestations in situ using technical instruments and laboratory analyses. This research, therefore, falls within a still underdeveloped field, as it integrates both remote sensing and field surveys, applying them to a strictly urban study area. Approximately twenty areas within Rome’s most densely urbanized zone, inside the Grande Raccordo Anulare, were deemed suitable for the study, with over 2,000 trees being georeferenced. The surveys allowed to document the health status of affected pines through a specifically structured evaluation sheet. Each tree was georeferenced and assessed based on the following qualitative parameters: canopy desiccation and density, presence and quantity of sooty mold, and overall tree vitality. Identifying a control area was crucial for correctly interpreting the results, allowing a clear distinction between infested and healthy trees. Given the spread of the infestation in Rome’s urban and peri-urban areas, the control area was selected in a pest-free pine forest within the Tenuta di San Rossore in Tuscany. In addition to field-collected information, remote sensing data were also acquired. Spectral values were extracted from PlanetScope images, provided by Planet Labs Inc., for the pixels corresponding to each georeferenced point. Then, using the spectral band values corresponding to red (Red) and near-infrared (NIR), the RDVI vegetation index was calculated. This index proved to be particularly useful in distinguishing differences in plant health conditions between areas. The analysis of the collected data was conducted on two levels: at the area scale, treating study zones as unique entities, and at the individual tree scale. The area-level analysis proved more effective in comparing healthy and infested zones, whereas tree-level observation presented challenges due to pixel variability, which was partially mitigated by the presence of multiple trees within the same area. The RDVI index trend was studied over 48 months across all surveyed areas, revealing a clear distinction between healthy and infested zones. During the period of greatest vegetative stress for P. pinea (summer), trees were severely weakened by the insect attack. Conversely, individual tree analysis was particularly useful for identifying a threshold to recognize infested trees. The results revealed a sharp difference 9 between infested and healthy trees: the identified threshold maximized classification accuracy based on the index. Finally, a comparison was made between satellite remote sensing data and field-acquired information. This comparison indicates that, in integrating the two data types, analyzing areas is methodologically more reliable than focusing on individual trees. Although satellite images offer sufficient resolution for many applications, evaluating the individual infestation level of trees is less accurate. Further methodological refinements and developments are needed to improve infestation monitoring. A crucial aspect concerns the more detailed collection of data on endotherapic treatments, as the absence of this information affects result interpretation. Additionally, remote sensing has intrinsic limitations: nadir images allow only canopy evaluation, making it difficult to detect damage to lower layers and distinguish canopy reflectance from background elements. Therefore, integrating remote sensing with field data remains essential and should be further developed and improved, building on the obtained results and addressing challenges and available resources.
Nel 2014, la cocciniglia tartaruga del pino (Toumeyella parvicornisCockerell) è stata rilevata per la prima volta in Italia, causando gravi infestazioni nelle pinete di pino domestico (Pinus pinea L.), soprattutto a Napoli e Roma. La presente ricerca si propone di: identificare i danni causati da T. parvicornis in alcune aree verdi della città di Roma; testare l’efficacia della scheda di valutazione messa a punto per il rilevamento delle infestazioni, valutare l’efficienza delle immagini dellacostellazione PlanetScope nel monitoraggio dell’impatto di T. parvicornis sui pini domestici (P. pinea); analizzare i vantaggi e le criticità dell’integrazione tra rilievi in campo e telerilevamento nell’individuazione, nel monitoraggio e nella gestione degli attacchi di insetti sulle alberature. Gli studi finora pubblicati sull’attacco della cocciniglia tartaruga al pino domestico sono stati condotti principalmente in alcune aree di Roma e Napoli. Tra questi, quelli basati sul telerilevamento si concentrano prevalentemente su zone suburbane o pinete estese, mentre gli altri monitorano le infestazioni in situ mediante strumentazione tecnica e analisi di laboratorio. Questa ricerca si inserisce quindi in un ambito ancora poco sviluppato, poiché integra sia il telerilevamento sia le indagini in campo, applicandoli a un’area di studio prettamente urbana. Sono state ritenute idonee per lo studio circa venti aree, nella zona più densamente urbanizzata di Roma, all’interno del Grande Raccordo Anulare, nelle quali oltre 2000 alberi sono stati georeferenziati. I rilievi hanno permesso di documentare lo stato fitosanitario dei pini colpiti, mediante una scheda di valutazione appositamente strutturata. Ogni pianta è stata georeferenziata e valutata secondo i seguenti parametri qualitativi: disseccamento e compattezza della chioma, presenza e quantità di fumaggine e vitalità generale dell’albero. Si è rivelata fondamentale l’individuazione di un’area di controllo per interpretare nella maniera più corretta i risultati ottenuti, consentendo di evidenziare in modo chiaro le differenze tra le piante attaccate e quelle sane. Data la diffusione dell’infestazione nella zona urbana e periurbana di Roma, la scelta dell’area di controllo è ricaduta su una pineta priva di infestazione, situata nella Tenuta di San Rossore in Toscana. Oltre alle informazioni raccolte in campo, anche tramite telerilevamento sono stati acquisiti dati. Dalle immagini PlanetScope, fornite dal servizio Planet labs. Inc., sono stati estratti i valori spettrali per i pixel corrispondenti a ciascun punto georeferenziato; successivamente con i valori delle bande spettrali corrispondenti al rosso (Red) e al vicino infrarosso (Nir), è stato calcolato l’indice di vegetazione RDVI. Questo è stato ritenuto particolarmente utile per discriminare le differenze, in termini di condizioni fitosanitarie, tra le aree. L’analisi dei dati raccolti è stata condotta su due livelli: a scala di area, trattando le zone di studio come entità uniche, e a scala di singolo albero. L'analisi a livello di area si è dimostrata più efficace nel confrontare zone sane e infestate, mentre l'osservazione delle singole alberature ha presentato delle criticità legate alla variabilità dei pixel, che è parzialmente compensata dalla presenza di più alberi all'interno di una stessa area. L’andamento dell’indice RDVI è stato studiato per un periodo di 48 mesi, per tutte le aree visitate, evidenziando una chiara separazione tra aree sane e infestate. In particolare, nel periodo di maggiore stress vegetativo del Pinus pinea (il periodo estivo), le piante vengono severamente debilitate dall’attacco dell’insetto. Al contrario, l’analisi delle singole alberature è stata particolarmente utile per identificare una soglia di RDVI per il riconoscimento delle piante attaccate. I risultati hanno rivelato una differenza netta tra piante infestate e sane: la soglia dell’indice individuata (33,1) massimizza l’accuratezza della classificazione sana/infestata. Infine, è stato compiuto un confronto tra i dati telerilevati da satellite e le informazioni acquisite in campo. Da tale confronto risulta che nell’integrazione tra le due tipologie di dati, analizzare le aree è metodologicamente più attendibile che un focus sulle singole piante. Sebbene infatti, le immagini satellitari offrano una risoluzione adeguata a molte applicazioni, risulta meno accurato valutare il grado di infestazione individuale degli alberi. Per migliorare il monitoraggio dell’infestazione, sono necessari ulteriori sviluppi e affinamenti metodologici. Un aspetto cruciale riguarda la raccolta più dettagliata dei dati sui trattamenti endoterapici applicati, poiché, la mancanza di questi, influenza l’interpretazione dei risultati. Inoltre, il telerilevamento presenta limitazioni intrinseche: le immagini nadirali permettono di valutare solo le chiome, rendendo difficile individuare i danni agli strati inferiori e distinguere la riflettanza della chioma da quella dello sfondo. Pertanto, l’integrazione con i dati rilevati in campo si conferma un’attività essenziale che va implementata e migliorata, partendo dai risultati ottenuti e valutando criticità e risorse.
Sviluppo di una metodologia di elaborazione di dati satellitari per il monitoraggio degli attacchi da insetti patogeni in ambiente urbano
FALANGA, Valentina
2025
Abstract
In 2014, the pine tortoise scale (Toumeyella parvicornis Cockerell) was detected for the first time in Italy, causing severe infestations in Pinus pinea L. forests, particularly in Naples and Rome. This research aims to: identify the damage caused by T. parvicornis in selected green areas of the city of Rome; test the effectiveness of the evaluation sheet developed for detecting infestations; assess the efficiency of images from the PlanetScope constellation in monitoring the impact of T. parvicornis on stone pines (P. pinea); and analyze the advantages and challenges of integrating field surveys and remote sensing in detecting, monitoring, and managing insect attacks on urban trees. The studies published so far on the attack of the pine tortoise scale on P. pinea have been conducted mainly in some areas of Rome and Naples. Among these, those based on remote sensing predominantly focus on suburban areas or extensive pine forests, while others monitor infestations in situ using technical instruments and laboratory analyses. This research, therefore, falls within a still underdeveloped field, as it integrates both remote sensing and field surveys, applying them to a strictly urban study area. Approximately twenty areas within Rome’s most densely urbanized zone, inside the Grande Raccordo Anulare, were deemed suitable for the study, with over 2,000 trees being georeferenced. The surveys allowed to document the health status of affected pines through a specifically structured evaluation sheet. Each tree was georeferenced and assessed based on the following qualitative parameters: canopy desiccation and density, presence and quantity of sooty mold, and overall tree vitality. Identifying a control area was crucial for correctly interpreting the results, allowing a clear distinction between infested and healthy trees. Given the spread of the infestation in Rome’s urban and peri-urban areas, the control area was selected in a pest-free pine forest within the Tenuta di San Rossore in Tuscany. In addition to field-collected information, remote sensing data were also acquired. Spectral values were extracted from PlanetScope images, provided by Planet Labs Inc., for the pixels corresponding to each georeferenced point. Then, using the spectral band values corresponding to red (Red) and near-infrared (NIR), the RDVI vegetation index was calculated. This index proved to be particularly useful in distinguishing differences in plant health conditions between areas. The analysis of the collected data was conducted on two levels: at the area scale, treating study zones as unique entities, and at the individual tree scale. The area-level analysis proved more effective in comparing healthy and infested zones, whereas tree-level observation presented challenges due to pixel variability, which was partially mitigated by the presence of multiple trees within the same area. The RDVI index trend was studied over 48 months across all surveyed areas, revealing a clear distinction between healthy and infested zones. During the period of greatest vegetative stress for P. pinea (summer), trees were severely weakened by the insect attack. Conversely, individual tree analysis was particularly useful for identifying a threshold to recognize infested trees. The results revealed a sharp difference 9 between infested and healthy trees: the identified threshold maximized classification accuracy based on the index. Finally, a comparison was made between satellite remote sensing data and field-acquired information. This comparison indicates that, in integrating the two data types, analyzing areas is methodologically more reliable than focusing on individual trees. Although satellite images offer sufficient resolution for many applications, evaluating the individual infestation level of trees is less accurate. Further methodological refinements and developments are needed to improve infestation monitoring. A crucial aspect concerns the more detailed collection of data on endotherapic treatments, as the absence of this information affects result interpretation. Additionally, remote sensing has intrinsic limitations: nadir images allow only canopy evaluation, making it difficult to detect damage to lower layers and distinguish canopy reflectance from background elements. Therefore, integrating remote sensing with field data remains essential and should be further developed and improved, building on the obtained results and addressing challenges and available resources.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_V_ Falanga.pdf
accesso aperto
Licenza:
Tutti i diritti riservati
Dimensione
5.03 MB
Formato
Adobe PDF
|
5.03 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/356249
URN:NBN:IT:UNIMOL-356249