PIECEWISE AND STRUCCHANGE: TEST OF TWO METHODSPIECEWISE AND STRUCCHANGE: TEST OF TWO METHODSOF CHANGE POINT ANALYSIS IN AGROCLIMATOLOGYOF CHANGE POINT ANALYSIS IN AGROCLIMATOLOGY

A.Chiaudani1, A. Berti2, M.Borin2, L.Mariani3

1. Veneto Region Environmental Protection Agency (ARPAV), Meteorological Centre of Teolo,

achiaudani@arpa.veneto.it

2. Padua University, Department of Environmental Agronomy and Vegetable production

antonio.berti@unipd.it; maurizio.borin@unipd.it

3. Milan University, Department of Vegetable production

meteomar@libero.it

Temperature9 stations: period 1956-2004

Precipitation49 stations: period 1956-2004

Historical data set 1956-2004

CMT Centro Meteorologico di Teolo Alessandro Chiaudani

p<0.01

CMT Centro Meteorologico di Teolo Alessandro Chiaudani

PIECEWISE BREAKPOINT

Piecewise regression: aims to identify the initial changing moment of the system, separating sub-periods, defined by their trends

CMT Centro Meteorologico di Teolo Alessandro Chiaudani

FLAT STEPS BREAKPOINT

Flat-step: aims to identify the transition from a climatic homogeneous phase to a new one (climatic normal), separating stationary sub-periods,

defined by their average values

CMT Centro Meteorologico di Teolo Alessandro Chiaudani

DIFFERENCE BETWEEN PEACEWISE AND FLATSTEPS

The piecewise regression breakpoint, identifying the beginning of the change, anticipates the flat-step

breakpoint that identifies the moment where the passage to a new stationary phase is demonstrable

CMT Centro Meteorologico di Teolo Alessandro Chiaudani

PHENOLOGY PROXY DATA

Other Proxy data

Annual anomaly of average Temp. in Italy (1961-2004) and

linear piecewise. Source: Toreti e Desiato (2007).

1981

CONCLUSIONS

The flat-steps approach can be particularly interesting for agro-climatic purposes, because it permits to identify the reference period for climatic normal which are the basis for a lot of evaluations about climatic change and adaptation of agricultural systems.