Forecasting the Future Drought Indices Due to the Effects of Climate Change in Al Najaf City, Iraq.

Drought is a natural disaster associated with a shortage of water availability for specified region within a specific time period. The impacts of drought are significant and extend to damage many important life aspects such as environmental, economic, and social activities. The forecasting of the drought events is an essential element for planning this disaster, reducing its effectiveness and response. The three characteristic frequency, intensity, and time period are the key parts for forecasting and assessment of droughts. Here, two drought indices (The Reconnaissance Drought Index (RDI), standardized precipitation index (SPI)) were used for forecasting of the future drought within Al Najaf city, Iraq. Thirty years meteorological data (average monthly precipitation and temperature) were used for the period (2021–2050) downloaded from the site of the Centre for Environmental Data Analysis (CEDA) for five grid points to cover overall study area. The computation of these indices conducted at a 12-month time scale and included the calculation of potential evapotranspiration by Thorthwaite method. The temporal drought intensity as well as drought frequency configurations were calculated and analyzed for each drought index. The results showed that the general average drought level expected will mildly dry while the maximum drought level expected will extremely dry. The more severe seasons of drought were forecasted in the years 2038, 2034 and 2021, respectively. Also, the prevailing event will be a one year drought and the maximum drought interval occurred within the study period will four consecutive years, with a 3.33% exceedance probability.


Introduction and Literature Review
Drought spell is a normal phenomenon that occurs as a result of the lowering values of rainfall depth than the normal limit [1]. Drought can be considered as one of most dangerous natural disasters, as its severe impacts on many human activities such as the agriculture, the environment, and the economy. Drought season developed over the long-term of time, this makes it difficult to define the start and end of this event, therefore, the drought is classified as an extraordinary hazardous. The dry spell within any hydrological system of a specific region is defined as a deficiency of available water [2]. In recent years, water resources have been subjected to increasing stresses as a result of several factors, the most important of which is the increase in population and the accompanying increase in the water demand from various domestically, agricultural, industrial and environmental sectors, in addition to the great shortage in water quantities as a result of climatic changes, and this has led to an increase in the international conflict over water shares. The Intergovernmental Panel on Climate Change (IPCC), 2007 indicated that in the near future, it will be approximately all areas of the world are subjected to negative impact of climate change on water resources, especially freshwater ecosystems. The durable strategic planning of water resources in any region is necessary in the face of the progressing climate change impacts, it is essential that these impacts be forecasted with a high level of spatial and temporal resolution [3]. A large number of researchers have studied climate change impacts on watersheds, only a few scientists have concentrated on watersheds management adaptation to the impacts of climate variability [4]. Over the previous period, climatologists and meteorologists have investigated the drought problem and developed a number of complex and simple drought indices such as Palmer Drought Severity Index and Precipitation Percentiles Index, respectively (WMO, 2012). The most communal index was the Standard Precipitation Index SPI, it can be considered as powerful and flexible tool established in 1993 by the American researchers, Doesken, McKee, and Kleist [5]. The SPI index is very easy in calculations, because the precipitation parameter is the only needed input independent parameter. In this study, average monthly precipitation and temperature data were downloaded from the site of the Centre for Environmental Data Analysis (CEDA) for five grid points covered overall study area for thirty year period (2021-2050) to assess the future drought indices. The RDI and SPI indices were computed for the next 30 years (2021-2050) in Al Najaf city, Iraq by using the DrinC software. The analysis of exceedance probability, frequency of spells of successive drought years, as well as reoccurrence period were conducted.

Study Region
Al Najaf is one of the important Iraqi cities, it is located southwest of the capital, Baghdad within the geographical coordinates of (29ᵒ 50' 00̋ -32ᵒ 21 00) N and (42° 50' 00 "-45° 44' 00") E, respectively, the Euphrates River passes it only through the eastern border [6][7]. This area is inhabited by more than one million citizens distributed over a number of main areas, the most famous of which are Al-Kufa, Al-Najaf central, Al-Abbasiyah, Al-Haidariyah, Al-Manathirah Al-Hurryah, and others [8]. The climate in Al Najaf is mainly dry to semi dry, its summer season hot with high temperature may be over 45°C in some days, while its winter is semi dry and cold with moderate temperature fluctuated between 8°C and 25oC and may reach to 0oC in some days. The average annual precipitation was assessed as 190.7 mm and 22.8 mm for wet and dry years, respectively. The season of rainfall extends between November (with moderate rainfall) and May (with intermittent rain showers) [9], figure 1 illustrated the geographical location for Al Najaf city.

Materials and Methodology
The methodology steps of this research focused on three stages are, pre-processing input data, executing DrinC program for calculating PET and drought indices, and post-processing of results, figure 2 explains the details of these steps.

3.1.Meteorological data (Expected precipitation and temperature)
For this forecasting study, A Representative Concentration Pathway model (RCP4.5) was selected for obtaining data of average monthly precipitation depths as well as average monthly temperature for 5 points were dispersed to covering the study area for the next 30 years (from 2021 to 2050). The site of the Centre for Environmental Data Analysis (CEDA) were used to downloaded these, this site available at: http://archive.ceda.ac.uk/. Figure 3 displays the distribution of these 5 points overall study area. Figures 4 and 5 indicate the precipitation and temperature input data which used to calculate the RDI and SPI indices.

3.2.1.Reconnaissance Drought Index (RDI).
RDI is considered as a new meteorological identification and assessment drought index, it was developed by [10] and presented in the MEDROPLAN coordinating meeting project. The computation of RDI based on the ratio between gathered amounts of precipitation and potential evapotranspiration Where: a CM represents the initial value of the RDI for the certain month, CM represents certain month during a year for a certain period, P j and PET j represent rainfall and evapotranspiration for the jth month within certain hydrological year, a ̅ CM is the average value of a certain month, y CM is the logarithm of a CM , y ̅ CM is the arithmetic mean of y CM , σ CM is the standard deviation [11]. Table 1, indicates the of meteorological drought intensity based on the SPI values. Table 2, showed the aridity zones limitation according to UNEP and FAO.  [12]. Previous studies have confirmed that the precipitation of short period irregularities affects the water content of the soil, whereas long period precipitation irregularities impact water resources [13]. SPI is characterized as applicable for evaluation of the severity of the droughts and helps to give early warnings of droughts. But, on the other hand, this index is depended only on the precipitation, therefore, its results can be considered rather weak. Simply, this index takes average precipitation as a reference line, then positive and negative values refer to up than and less than compared with average precipitation [14], Table 3, indicates the of meteorological drought intensity based on the SPI values. The main distribution function which used to give a time series of precipitation is the function of gamma probability, it is computed as follows [16]: If considered that the function G(p) represent the quantity of precipitation in mm, γ and α represent the shape parameters, and Γ (α) is the Gamma function of α. p ̅ represent the average quantity of precipitation in mm, n represent data number of precipitation.
Where possible, in arithmetic standardizing the data directly from a fitted natural distribution, then SPI is considered as the next form: = − (17)

3.3.Calculations of potential evapotranspiration (PET)
Using of any potential evapotranspiration calculation method does not give the impression to affect the RDI results. There are many methods to calculate PET such as Penman Monteith, Blaney-Criddle (Doorenbos & Pruitt 1977) [17], Thornthwaite (1948) [18], and Hargreaves andSamani (1982, 1985) [ [19][20], The first two methods require minimum and maximum temperature records, whereas the third and fourth method requires only mean temperature records. Here, the PET was calculated based the Thornthwait Method as follows: Where: M represents the mean monthly temperature in °C units, N represents the number of monthly measurements, j is the index of annual heat (°C), i is the monthly temperature coefficient (°C), D is the number of the days a month, S is the average number of sunshine hours. Figure 6 showed the expected PET for five points within future periods from 2021 to 2050.

4.1.RDI Results
To obtain clearer representation of spells of wets and droughts, RDI values were implemented for 12 months to cover the yearly precipitations of the points during a year. Figures 7-9, illustrate the three RDI types values for the next 30 years, figure 10, indicates the RDI values for annual average precipitation and temperature data. As can be seen from these figures, the region will suffer for sixteen years of droughts, the ratio of wet years was about 47% while for the dry years about 53%. According to the index divisions in table (1), the RDI levels fluctuated from extremely dry to extremely wet. Generally, the calculations referred to that the average aridity value=0.08, and this means that the climate in this region is arid according to limitations of table (2). The first part of the study period (2021-2035) can be considered a wet period, for the reason that it has nine wet years besides six dry years, while the second part of the period (2036-2050) is a drought period, because it has ten dry years with only five wet years. Compared to the figure 6, the expected PET will increase during the first part of the study period and it reaches its peak in the second part. Figure 11, expresses the replication of the successive dry years in addition to the exceedance probability percentage with frequency. This figure indicates that the repetition of the two successive dry years happened nine spells and is concentrated in the second part of the study period (2036-2050). Moreover, it can be noticed that the repetition of one drought years was most frequent recurrence, it occurs with exceedance probability of 53%, followed by two successive drought years with an exceedance probability of 30%. The maximum drought period can be occurred within the study period was four successive years, with a 3.33% exceedance probability.

4.2.SPI Results
As in the previous procedure of calculations, SPI values were depended on 12 months to cover the yearly precipitations of the points during a year. Figure 12, illustrates the SPI values for the next 30 years, figure 13, indicates the SPI values for annual average precipitation data. The SPI level for the study region ranges from extremely wet to extremely dry. There are fourteen dry years as opposed to sixteen wet years, this means that the ratio of dry years is 47% and the ratio of wet years is 53%. As shown in the previous index (RDI), the SPI results indicate that there are two parts, the first part of the study period (2021-2035) can be described as a wet period due to that it has ten wet years besides five dry years, while the second part of the period (2036-2050) is a drought period, because it has six wet years with nine dry years. Figure 13, indicates that strong fluctuations will occur in the period from 2026 to 2041. Figure 14, expresses the repetition of the successive dry years in addition to the exceedance probability percentage with frequency for SPI index. This figure illustrates that the repetition of the two successive dry years occurred nine times and is concentrated in the second part of the study period (2036-2050). In addition, it can be noticed that the repetition of one drought years was most frequent recurrence, it occurs with exceedance probability of 47%, followed by two successive drought years with an exceedance probability of 23%. The maximum drought interval can be occurred within the study period was four consecutive years, with a 3.33% exceedance probability.