Functional data analysis of temperature and rainfall observations for the regional indexing of climate change severity in india (PhD)

Abstract

The persistence of global warming and its impact on various facets of human life is indubitable. Particularly, India is among the countries which are most vulnerable to the influence of climate change. Thus, identification of spatial and temporal course of climate change events in India is crucial to develop the optimal adaption and mitigation strategies. The present work aims to identify the footprints of temperature change in Indian context by analyzing monthly temperature data pertaining to the period of 1901-2005. Functional data analysis (FDA) has been implemented for the detection of mean change-years in Indian temperature. The technique is more robust than conventional multivariate analysis and provides more information about the data. The estimated change-years have been observed to largely pertain to the latter half of 20th century and fall in close proximity of widely reported global change-years. The all-India temperature shift is estimated to be 0.53◦C during 1901-2005. An index, consolidating the results of FDA such as change-years and temperature rise, has been developed to quantify the temperature change severity. Based on the spatial distribution of index values, the hot-spots have been found to cluster over the north-central and eastern parts of the country. These findings have been validated against actual conditions. The approach has also been employed to understand the manifestation of significant changes in maximum and minimum temperatures. It is observed that maximum temperatures have undergone frequent changes during the period 1901-1970, whereas continual changes are observed in minimum temperatures during post 1970. The values obtained from the developed index suggests that change in minimum temperature is more severe and wide spread than change in maximum temperature. Further, changing patterns of climate have substantially altered the frequency of different rainfall events. To investigate the issue, firstly regional rainfall observations have been converted into smooth curves using FDA. Subsequently, wet spells for Indian homogeneous regions have been estimated from their corresponding smooth curves. Thereafter, decadal distribution of moderate, heavy and extremely heavy rainfall events have been computed during the estimated wet spells. It has been found that frequency of extremely heavy rainfall events have increased in every region due to the warmer climate.Buildings are fundamental components of infrastructure and play a crucial role in the dynamics of climate change. To assess the joint impact of wind and rain endured by vertical building facades, annual driving rain index (aDRI) is computed on gridded scale. The aDRI values have been estimated using gridded daily records of wind and rainfall pertaining to the sixty-year period of 1951-2010. The study compares the quality of aDRI estimates as a function of the temporal resolution of gridded data by evaluating total, non co-occurring and averaging errors. An aDRI map for India at 1◦ ×1◦ (lat./long.) resolution has been constituted and the dominant directions of driving rain have been identified. The discourse concludes with a trend analysis of yearly driving rain index values.