The availability of high-resolution datasets describing the spatio-temporal evolution of precipitation is becoming increasingly important in order to analyse the long-term variability and trends of the climatic signal and possible impacts over specific areas of interest. These datasets are therefore crucial not only for research purposes, but also for decision-makers in a wide range of fields, such as agriculture, energy production, hydrology, natural risk monitoring and resource management. The reconstruction of high-resolution climate descriptions requires both accurate in-situ observations and suitable interpolation schemes to project the station data onto regular grids. The study focuses on the development and improvement of interpolation methods for monthly precipitation data and on the reconstruction and analysis of gridded datasets from dense and high-quality rain-gauge records covering specific study domains. The first goal of the work was the reconstruction of the high-resolution monthly precipitation climatologies over Italy for the 1961-1990 period. The observation database was set up thanks to a relevant effort of data rescue and collection from a great number of national and international sources and led to more than 4500 quality-checked monthly records available for the climatological reconstruction. Considering the heterogeneous Italian orography and the influence of surface features on precipitation distribution, a local weighted linear regression (LWLR) of precipitation versus elevation was applied to interpolate the station monthly normals onto a 30-arc second resolution Digital Elevation Model (DEM). Leave-one-out (LOO) validation and inter-comparison proved that the approaches modelling the local precipitation-orography relationship provide more accurate results in respect with methods considering larger spatial scales or not including the topography at all. The computed 30-arc second resolution monthly precipitation climatologies for Italy were on-line released and they represent an updated and highly detailed description of precipitation normal over the whole national territory in digital form. In order to evaluate the temporal evolution of precipitation over some of the most vulnerable Italian regions, gridded datasets of long-term precipitation series were produced by means of the anomaly-based approach. The secular precipitation series were reconstructed for the upper part of Adda river basin (Central Italian Alps), with an additional focus over the Forni glacier, and for Sardinia, as case-study for the Mediterranean area. New records were collected, especially thanks to the integration of the recent automatic station records with those of the previous mechanical networks and to digitisation activities of the most ancient data from hardcopy archives. All the series underwent statistical procedures aiming at avoiding inhomogeneities due to non-climatic signals. The gridded dataset allowed to get the secular areal precipitation records for the study regions and to evaluate their trends. As regards the Adda basin, the reconstructed 1800-2016 series showed statistically significant negative trends for annual and autumn precipitation. The comparison with the 1845-2016 annual basin runoff record, which is one of the longest runoff series available in Italy, allowed both to depict a strong decrease in annual runoff driven by the increase of evapotranspiration and to evaluate the possible contribution of gauge undercatch, especially in mountainous sites, to the underestimation of basin precipitation. The trend analysis over the 1922-2011 areal monthly precipitation record computed for Sardinia highlighted statistically significant decreases of -2.3% and -4.1% decade-1 in annual and winter values, respectively, and a positive but not statistically significant trend in summer precipitation. These outcomes agreed with other literature studies focusing on precipitation variability over Mediterranean area. The final part of the work focused on the development of interpolation schemes to improve the accuracy of gridded precipitation fields for domains unevenly covered by station networks. Norway represents a very interesting case-study, where the severe climatic conditions and the complex orography limit the management of in-situ observations over the most remote regions leading to an unbalanced station distribution between North and South and between low and high elevation. At this aim, a new method to compute the Norwegian monthly precipitation climatologies (1981-2010) at 1 km resolution was implemented and the gridded dataset was on-line released for both research and operative purposes. In this scheme, named HCLIM+RK, the HCLIM-AROME climate numerical model fields, which are not based on rain-gauge data and describe the precipitation gradients also over unsampled areas, are combined with available in-situ observations by the kriging interpolation of station residuals. In HCLIM+RK the high-resolution reliable precipitation patterns provided by the numerical model are retained and the station data are used to correct the biases affecting the numerical fields. The LOO reconstruction errors of Norwegian station normal showed that the combined approach almost removes the biases affecting the original HCLIM-AROME dataset and it provides much lower errors than conventional interpolation procedures based on stations only.

PRECIPITATION CLIMATOLOGIES AND LONG-TERM RECORD RECONSTRUCTION: STUDIES IN A CHANGING CLIMATE / A. Crespi ; scientific tutor: M. Maugeri; scientific co-tutor: M. Brunetti. - Milano : Università degli studi di Milano. DIPARTIMENTO DI SCIENZE AGRARIE E AMBIENTALI - PRODUZIONE, TERRITORIO, AGROENERGIA, 2019 Jan 30. ((31. ciclo, Anno Accademico 2018.

PRECIPITATION CLIMATOLOGIES AND LONG-TERM RECORD RECONSTRUCTION: STUDIES IN A CHANGING CLIMATE

CRESPI, ALICE
2019-01-30

Abstract

The availability of high-resolution datasets describing the spatio-temporal evolution of precipitation is becoming increasingly important in order to analyse the long-term variability and trends of the climatic signal and possible impacts over specific areas of interest. These datasets are therefore crucial not only for research purposes, but also for decision-makers in a wide range of fields, such as agriculture, energy production, hydrology, natural risk monitoring and resource management. The reconstruction of high-resolution climate descriptions requires both accurate in-situ observations and suitable interpolation schemes to project the station data onto regular grids. The study focuses on the development and improvement of interpolation methods for monthly precipitation data and on the reconstruction and analysis of gridded datasets from dense and high-quality rain-gauge records covering specific study domains. The first goal of the work was the reconstruction of the high-resolution monthly precipitation climatologies over Italy for the 1961-1990 period. The observation database was set up thanks to a relevant effort of data rescue and collection from a great number of national and international sources and led to more than 4500 quality-checked monthly records available for the climatological reconstruction. Considering the heterogeneous Italian orography and the influence of surface features on precipitation distribution, a local weighted linear regression (LWLR) of precipitation versus elevation was applied to interpolate the station monthly normals onto a 30-arc second resolution Digital Elevation Model (DEM). Leave-one-out (LOO) validation and inter-comparison proved that the approaches modelling the local precipitation-orography relationship provide more accurate results in respect with methods considering larger spatial scales or not including the topography at all. The computed 30-arc second resolution monthly precipitation climatologies for Italy were on-line released and they represent an updated and highly detailed description of precipitation normal over the whole national territory in digital form. In order to evaluate the temporal evolution of precipitation over some of the most vulnerable Italian regions, gridded datasets of long-term precipitation series were produced by means of the anomaly-based approach. The secular precipitation series were reconstructed for the upper part of Adda river basin (Central Italian Alps), with an additional focus over the Forni glacier, and for Sardinia, as case-study for the Mediterranean area. New records were collected, especially thanks to the integration of the recent automatic station records with those of the previous mechanical networks and to digitisation activities of the most ancient data from hardcopy archives. All the series underwent statistical procedures aiming at avoiding inhomogeneities due to non-climatic signals. The gridded dataset allowed to get the secular areal precipitation records for the study regions and to evaluate their trends. As regards the Adda basin, the reconstructed 1800-2016 series showed statistically significant negative trends for annual and autumn precipitation. The comparison with the 1845-2016 annual basin runoff record, which is one of the longest runoff series available in Italy, allowed both to depict a strong decrease in annual runoff driven by the increase of evapotranspiration and to evaluate the possible contribution of gauge undercatch, especially in mountainous sites, to the underestimation of basin precipitation. The trend analysis over the 1922-2011 areal monthly precipitation record computed for Sardinia highlighted statistically significant decreases of -2.3% and -4.1% decade-1 in annual and winter values, respectively, and a positive but not statistically significant trend in summer precipitation. These outcomes agreed with other literature studies focusing on precipitation variability over Mediterranean area. The final part of the work focused on the development of interpolation schemes to improve the accuracy of gridded precipitation fields for domains unevenly covered by station networks. Norway represents a very interesting case-study, where the severe climatic conditions and the complex orography limit the management of in-situ observations over the most remote regions leading to an unbalanced station distribution between North and South and between low and high elevation. At this aim, a new method to compute the Norwegian monthly precipitation climatologies (1981-2010) at 1 km resolution was implemented and the gridded dataset was on-line released for both research and operative purposes. In this scheme, named HCLIM+RK, the HCLIM-AROME climate numerical model fields, which are not based on rain-gauge data and describe the precipitation gradients also over unsampled areas, are combined with available in-situ observations by the kriging interpolation of station residuals. In HCLIM+RK the high-resolution reliable precipitation patterns provided by the numerical model are retained and the station data are used to correct the biases affecting the numerical fields. The LOO reconstruction errors of Norwegian station normal showed that the combined approach almost removes the biases affecting the original HCLIM-AROME dataset and it provides much lower errors than conventional interpolation procedures based on stations only.
MAUGERI, MAURIZIO
precipitation climatology; interpolation; variability and trends
Settore FIS/06 - Fisica per il Sistema Terra e Il Mezzo Circumterrestre
PRECIPITATION CLIMATOLOGIES AND LONG-TERM RECORD RECONSTRUCTION: STUDIES IN A CHANGING CLIMATE / A. Crespi ; scientific tutor: M. Maugeri; scientific co-tutor: M. Brunetti. - Milano : Università degli studi di Milano. DIPARTIMENTO DI SCIENZE AGRARIE E AMBIENTALI - PRODUZIONE, TERRITORIO, AGROENERGIA, 2019 Jan 30. ((31. ciclo, Anno Accademico 2018.
Doctoral Thesis
File in questo prodotto:
File Dimensione Formato  
phd_unimi_R11342.pdf

accesso aperto

35.57 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

Caricamento pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/2434/614284
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact