Non-stationarity Detection And The Driving Forces Exploration Of The Precipitation-temperature Dependency Structure Under Changing Environment

In recent decades,the climate has changed in the Loess Plateau obviously.At the same time,one of the typical human activities——Afforestation/Conversion of Farmland to Forest(Grassland)has also changed the underlying vegetation significantly.Under the changing environmrent,precipitation and temperature is the most important and most intuitive response climate index to climate change.With the dual effects of climate change and underlying surface change togetwher,how the relationship between climate elements changes and how the response process exists with vegetation growth and grain production is of great significance for understanding the response process of hydrological cycle on the Loess Plateau.The loess plateau was selected as a typical area in this article,both precipitation and temperature were selectes as the main research elements;climate joint distribution were constructed by copula methed,the non-stationarity of the dependency structure was detected via the copula-based likelihood-ratio method,also,double mass curve was introduced to verify the accuracy of the result.Based on this,the characteristics of the dependency structure before and after the mutation and the whole series wereanalyzed.Random forest method and projection importance criterion of variables were used to select the main driving forces of the dependency structure quantitatively.Finally,ridge regression method was introduced to establise the driving force regression equation of dependency structure,and the driving mechanism of dependent structure change was explored by correlation analysis.The results showed that:(1)There are more change points of the precipitation-temperature dependency structure than the temperature-temperature dependency structure between two adjacent season on the Loess Plateau,which means the consistency of the precipitation-temperature dependency structure is more sensitive to the climate change;(2)The main driving force of the dependency structure are the local climatic factors(precipitation,air temperature,wind speed,vapor pressure,relative humidity,sunshine time,etc.),it is important to note that the NDVI is also one of the main driving forces of some dependency structure:summer precipitation-the lowest average temperature in summer,summer precipitation-average temperature in summer,the average minimum temperature in summer-average minimum temperature in autumn,average temperature in summer-average temperature the autumn;(3)Correlation analysis between the local meteorological factors and teleconcordance factors(AO,PDO,ENSO and Sunspots)proved that there was a significant correlation between PDO and air temperature and wind speed,ENSO and relative humidity,sunshine hours and Sunspots,etc.,which provided a premise for further exploration of the driving mechanism of dependent structure.Generally,this study provides important insights into the response of the P-T dependency structure dynamics to a changing environment,where the proposed research framework could be extended to any other watershed and any bivariate hydrometeorological elements.