Identifying hotspots cities vulnerable to climate change in Pakistan under CMIP5 climate projections

Abstract

In this study, an ensemble of statistically downscaled 14 multi-global climate models for RCP4.5 and RCP8.5 emission scenarios was employed to implement a comprehensive assessment of climate change impacts over Pakistan in order to identify the future hotspots cities in terms of changes in temperature and precipitation. The analyses focused on the minimum, maximum and average temperature and precipitation in three time-slices: 2006-2035, 2041-2070, and 2071-2,100. Average temperature is projected to increase by 2.6 C under RCP4.5 while 5.1 C under RCP8.5 by the end of this century with the north side of Pakistan (mainly over North Pakistan-NP, Monsoon Region-MR and Khyber Pakhtunkhwa-KP) presenting the highest changes in the temperatures. Wetter conditions are expected in the future over Pakistan, mainly over the MR. In general, air temperature and precipitation showed linear positive correlation over Pakistan in both RCP scenarios. Hotspot cities where extreme climate, that is, the hottest, dryer and wetter, exists were also identified. Hyderabad will likely become the hottest city of Pakistan by end century with the highest average temperature reaching 29.9 C under RCP4.5 and 32.0 C under RCP8.5 followed by Jacobabad, Bahawalnagar, and Bahawalpur. Most of the hottest cities are detected in areas on the southern side of Pakistan. On the other hand, the wettest cities, Murree, Balakot and Muzaffarabad, are located in the MR. Dry conditions are likely to be prevalent in Dalbandin followed by Khanpur and Jacobabad under both RCPs. The uncertainties of the projections were also evaluated. For precipitation, for example, there are a large number of outliers indicating the high variability/uncertainties of the projections. These uncertainties are clearer when the probability density functions are analysed for individual sub-domains in Pakistan.

Temperature Projections over the Indus River Basin of Pakistan Using Statistical Downscaling

“…A combination of increased surface albedo [7], cloud radiative forcing, e.g., [75], and soil moisture feedbacks can (at least partly) explain a more significant T min warming within the UIB. A projected increase in WS precipitation, which is robust across many studies, e.g., [40,63,76], further supports such feedbacks. Since the WS precipitation mostly falls as snow, increased albedo from the fresh snow surface (and associated cloud covers) may largely explain a smaller increase in T max .…”

“…Moreover, the magnitude of projected warming differs significantly among different studies. For instance, a recent study projected about 6 • C warmer UIB [40], but nearly half of this warming is reported by [41] under similar radiative forcing. Considering greater sensitivities of the Indus flows towards warming magnitude, where 1 • C rise in the mean temperature can increase up to 16% glacial flows [42], such warming discrepancies can seriously implicate the adaptation planning.…”

“…[43] argue that the basin characterization using climate patterns rather than arbitrarily defined sub-regions, e.g., refs. [36,40] can support a more realistic climate analysis.…”

“…[55] may not accurately correct such systematic biases, particularly over high-altitude regions. Similarly, the uncertainty analysis using GCM simulated temperatures and insufficient observations may lack fidelity, e.g., [40]. In contrast, the GCMs' ability to simulate atmospheric circulation dynamics has considerably improved over time, e.g., [56,57].…”

Precipitation projections over the Indus River Basin of Pakistan for the 21st century using a statistical downscaling framework

“…Besides, our ensemble also includes their secondary considerations for Dry-Warm (CMCC-CMS) and Wet-Cold (CNRM-CM5) regional future (see Table 2). Similarly, the 14-model CMIP5 ensemble of Ali et al (2020) also contains almost all our models. Such ensemble similarities justify the regional relevance of our models, despite its smaller sample size.…”