Development of drought intensity-duration-frequency curves based on standardized climatic indices using physical variables, deficit in precipitation and deficit in streamflow

Abstract

Drought is a natural disaster that causes water scarcity and affects the population seriously in many ways such as economic, social and environmental impacts. These effects have been further aggravated by rising demands for agricultural, municipal and industrial water, etc. due to the growing population coupled with rising life standards and increasing industrialization in the world. Compared to other natural hazards, drought is among the most damaging disasters as it develops slowly over a considerable period of time that may linger for several years even after the drought has terminated. How important the drought is in the sustainability of water resources requires an extra effort to understand and explain. Domestic use, agriculture, industry, energy, tourism and many more sectors are exposed to drought, which has a direct impact on ecology, economy and society. As one of the least understood hydrological phenomena, drought stands as both a technical and social problem. This is a great motivation to explore drought concepts and methodologies in this thesis by extending them from their quantitative technical context to qualitative social context. Case studies were performed in two river basins with different climatic characteristics. One is Seyhan River Basin in the south of Turkey with a Mediterranean climate, and another is the Kocher catchment in the southwestern corner of Germany with a humid subtropical climate. In the literature, the drought is commonly analyzed by the use of available hydroclimatic or hydrologic data with little in-depth consideration of specific major dry periods experienced over a region. Also, it is not a common practice to assess the probability of drought categories with a rolling time series and hence the changing knowledge for operational drought monitoring. A combination of such quantitative analysis with a comprehensive qualitative assessment of drought as a human-water relation aimed to fill this gap by performing a case study in the Seyhan River Basin, Turkey. Therefore, the first part of this thesis focuses on non-stationary rolling time series analysis together with the stationary full-record time series analysis and the individual major dry period analysis. Six major dry periods were identified from the precipitation time series of 19 meteorological stations. Dry periods extended over 1970-1974, 2002-2008, and 2013-2015, known also as droughts of 1974, 2008, and 2014, respectively, should particularly be noted among all as they affected several sectors in the river basin with the long-lasting deficit in water. Major dry periods were analyzed by rolling time series and full time series, and they were also analyzed individually. The results show that significantly higher probabilities were calculated for extreme droughts with the use of individual major dry periods. An important outcome of the study is that drought is underestimated in practice with the sole use of the whole data record. Also, a major dry period could be important in terms of its duration while another in terms of its severity or intensity. Each major dry period has its own impact on human-water relations that can be influential on drought mitigation, management and governance. The second part of the thesis investigates the statistical properties of drought characteristics for a better understanding of drought to mitigate its negative impacts and improve drought management strategies. Differently from the common approach based on the run theory, the hypothesis is that drought is different from the dry period and dry periods identify droughts of different durations. Up to now, no attention has been paid in this sense to the statistical properties of dry periods versus those of more severe and impactful droughts as separate phenomena. The study shows how different the statistical properties of dry periods and droughts are by quantifying them with the Standardized Precipitation Index (SPI) together with the total probability theorem-coupled frequency analysis. The study tests four concepts in this context: (1) Dry period, (2) Critical drought for a dry period, (3) Critical drought for a year. The results uncover the high number of droughts, which are not addressed in the run theory, and reveal more intense droughts that are masked within the dry periods. Specifically, the drought impact is more often related to drought intensity than severity and duration. In addition, the dry period is less intense than other concepts. Among the droughts newly proposed, critical droughts are the most intense concepts. This demonstrates that drought characteristics calculated for the drought concepts are different from the common approach of the dry period. The difference among the concepts is large enough to carry potential implications for local or regional scale drought management planning. In the third study, drought intensity-duration-frequency (IDF) curves were developed by using the critical drought identified in the second study. The drought IDF curves are based on precipitation and streamflow deficits because drought estimates in terms of physically measurable variables are key knowledge for effective water management. However, how these deficits vary with the drought event severity indicated by commonly used standardized indices is often unclear. Drought characteristics assigned the same value in the index are not necessarily the same in different regions, and in different months of the same region. The study investigates drought to remove this disadvantage of the index-based drought IDF curves and develop intensity-duration-frequency (IDF) curves in terms of the associated deficit. In order to study the variation of deficits, the link between precipitation and streamflow was used, and the associated indices, standardized precipitation index (SPI) and standardized streamflow index (SSI). More specifically, the analysis relies on frequency analysis combined with the total probability theorem applied to the critical drought severity. The critical drought has varying durations and it is extracted from dry periods. IDF curves in terms of precipitation and streamflow deficits for the most severe drought of each drought duration in each year are then subject to the comparison of statistical characteristics of droughts for different return periods. Precipitation and streamflow data from two catchments, the Seyhan River (Turkey) and the Kocher River (Germany) provide examples of two climatically and hydrologically different cases. A comparison of the two cases allows to test a similar method in different hydrological conditions. The results show that precipitation and streamflow deficits vary systematically reflecting seasonality and the magnitude of precipitation and streamflow characteristics of the catchments. Deficits change from one month to another at a given station. Higher precipitation deficits were observed in winter months compared to summer months. Additionally, major droughts experienced in both catchments on the IDF curves show that the major droughts have return periods at the order of years at short durations. This coincides with the observation in the catchments and shows the applicability of the IDF curves. The IDF curves can be considered a tool for the use in a range of specific activities of agriculture, ecology, industry, energy, water supply, etc. This is particularly important to end-users and decision-makers to act against the drought quickly and precisely in a more physically understandable manner. Other than quantitative methodologies, drought was also analyzed by a qualitative approach. Using the causal loop diagram, the drought event of 2008 observed in the Seyhan River Basin was taken as an example to focus on the identification of bidirectional feedback between humans and water under drought conditions. This particular drought is short but severe as it coincided with the least amount of precipitation ever recorded in the river basin. It is an important example because it gave an impetus to the governmental water authorities to initiate basin-scale drought management plans. The quantitative and qualitative data collected for this drought were used in the causal loop analysis. The causal loop diagram shows that short-term measures under economic, structural and conservation frameworks can only provide temporary solutions which may aggravate drought impact. A sustainable water availability under drought conditions can only be achieved by long-term reactions such as drought management plans. In summary, the thesis contributes to drought analysis research by proposing new methodologies and drought concepts. The outcomes of the study underline the need for such drought analyses. The analysis of the conceptualization reveals strong differences among the statistical properties of the drought characteristics. Moreover, the thesis emphasizes the region-specific characteristics of drought that need to be considered for drought management strategies.