Journal of Spatial Analysis

Understanding the changes in extreme precipitation over a region is very important for adaptation strategies to climate change. One of the most important topics in this field is detection and attribution of climate change. Over the past two decades, there has been an increasing interest for scientists, engineers and policy makers to study about the effects of external forcing to the climatic variables and associated natural resources and human systems and whether such effects have surpassed the influence of the climate’s natural internal variability. The definitions used in the 5th assessment report were taken from the IPCC guidance paper on detection and attribution, and were stated as follows: “Detection of change is defined as the process of demonstrating that climate or a system affected by climate has changed in some defined statistical sense without providing a reason for that change. An identified change is detected in observations if its likelihood of occurrence by chance due to internal variability alone is determined to be small.  Attribution is defined as the process of evaluating the relative contributions of multiple causal factors to a change or event with an assignment of statistical confidence”. Detection and attribution of human-induced climate change provide a formal tool to decipher the complex causes of climate change. In this study the optimal fingerprinting detection and attribution have been attempted to investigate the changes in the annual maximum of daily precipitation and the annual maximum of 5-day consecutive precipitation amount over the southwest of Iran.

This is achieved through the use of the Asian Precipitation—Highly Resolved Observational Data Integration Towards Evaluation of Water Resources Project(APHRODITE) dataset as observation, a climate model runs and the standard optimal fingerprint method. To evaluate the response of climate to external forcing and to estimate the internal variability of the climate system from pre-industrial runs, the Norwegian Climate Center’s Earth System Model- NorESM1-M was used. We used up scaling to remap both grid data of observations and simulations to a large pixel. This remapped pixel coverages the area of the southwest of Iran. The optimal finger printing method needs standardized values like probability index(PI) or anomalies as input data, since the magnitude of precipitation varied highly from one region to another. The General Extreme Value distribution (GEV) is used to convert time series of the Rx1day and Rx5day into corresponding time series of PI.  Then we calculated non-overlapping 5-year mean PI time series over the area study. In this research, we applied optimal fingerprinting method by using empirical orthogonal functions.  The implementation of optimal fingerprinting often involves projecting onto k leading EOFs in order to decrease the dimension of the data and improve the estimate of internal climate variability. A residual consistency test used to check if the estimated residuals in regression algorithm are consistent with the assumed internal climate variability. Indeed, as the covariance matrix of internal variability is assumed to be known in these statistical models, it is important to check whether the inferred residuals are consistent with it; such that they are a typical realization of such variability. If this test is passed, the overall statistical model can be considered suitable.

Results obtained for response to anthropogenic and natural forcing combined forcing (ALL) for Rx1day and Rx5day show that scaling factors are significantly greater than zero and consistent with unit. These results indicate that the simulated ALL response is consistent with Rx1day observed changes. Also, it is found that the changes in observed extreme precipitation during 1951-2005 lie outside the range that is expected from natural internal variability of climate alone and greenhouse gasses alone, based on NorESM1-M climate model. Such changes are consistent with those expected from anthropogenic forcing alone. The detection results are sensitive to EOFs. We estimate the anthropogenic and natural forcing combined attributable change in PI over 1951–2005 to be 1.64% [0.18%, 3.1%, >90% confidence interval] for RX1day and 2.5% [1%,4%] for RX5day.