EVALUATION OF DROUGHT IN IRAQ USING DSI . BY REMOTE SENSING

Evaluation of drought patterns in Iraq and determining the most susceptible areas of this phenomenon were analyzed, using the remotely-sensed Drought Severity Index (DSI) through analysis the daily and annual DSI for three zones over Iraq, also have been analyzed DSI time series using run theory to evaluate the characteristics of drought in Iraq. The efficiency of DSI for drought monitoring was examined from compared with Percentage of Precipitation Anomaly (PPA) for three zones (Arid and Semi-Arid, Steppes and Desert), and compared with drought indicators (Evapotranspiration (ET), Potential evapotranspiration (PET) and total annual precipitation (PRE)) for the period 2000-2011, were derived from the Numerical Terradynamic Simulation Group (NTSG). The spatial interpolation techniques in Geographic Information System (GIS) package has been used, to cover the whole extent of country and extracting the zones. Statistical methods were applied to compute the probability of drought events at every zone. The results showed the drier year is 2008, the wetter years are 2001 in Desert zone and 2003 in steppes and Arid and Semi-Arid Zone zones. The results also showed a significant fluctuation in precipitation from the average, especially at Arid and Semi-Arid Zone when compared with other zones. The values of standard deviation of precipitation were compared with precipitation anomalies for each zone, Arid and Semi-Arid is the drier zone in 2007-2008, the wetter zone is also Arid and Semi-Arid in 2002-2003. Using run theory, the drier Zone is Arid and Semi-Arid and the wetter Zone is steppes during study period.


INTRODUCTION
Drought is a random natural phenomenon that emerges from a large deficiency in precipitation.It is the costliest natural disasters in the world and affects a very large number of people every year, it is important to monitor, understand and may be predict their frequency (21).The causes for the incidents of drought are complex because they don't just depend on the atmosphere but also on the hydrologic processes which provide moisture to the atmosphere (4).Drought events can be determined by characteristics of drought using the runs theory.Yevjevich in 1967 (23) who is proposed the runs theory and defined droughts as periods during which the water supply does not meet the current water demand by determining drought variables and their drought characteristics: (a) duration, (b) severity, and (c) intensity from the streamflow time series (12).This method has been widely used in the field of hydrology and there is a large amount of research on the characterization of droughts and the first few works that applied the runs theory in hydrology by (24), ( 18), ( 19), (11), (8), and (20) and has been applied in several drought models and analyzes, (for example, ( 6) and (13), which can be estimated the return periods of extreme events (7).This study focuses on five objectives: (i) Evaluation of the behavior of drought Indicators used in this study, (ii) Evaluation of the efficiency of remotelysensed Drought Severity Index (DSI) for drought monitoring by Compared with Percentage of Precipitation Anomaly (PPA), (iii) Determining drought events by using DSI through analysis of daily and annual mean of DSI data for the period (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011) (14).Numerical Terradynamic Simulation Group (NTSG) calculated DSI by using MODIS ET/PET and MODIS NDVI (9), which use global earth's data derived from NASA satellites (Terra and Aqua).The DSI dataset for the whole of earth 109.03 million km 2 of the vegetation areas of the earth's surface for the period (2000-2011) at 8-day and annual time periods, and spatial resolution 0.5 and 0.05 degrees (17).

Rainfall Data
In this study, real data of monthly rainfall (PRE) from ground stations were obtained from the Iraqi Meteorological Organization and Seismology (IMOS) for the period 1980-2015.The dataset was collected from 39 as shown in figure (1).The database was digitally encoded into a Geographic Information System (GIS) database for a distribution map which shows the general distribution of seasonal rainfall and climatic zones for the study area (3).
Table 1.Data Sources  1) and (10).Rainfall data collected from weather stations for each rainfall season (October, November, December, January, February, March, April and May), to get total seasonal rainfall.Then, were aggregated to the Zones (Arid and Semi-Arid, Steppes, and Desert) to get the average of each Zone.

Fig. 1. Weather Stations and Climatic Zones Methods
Remotely-Sensed DSI: The ratio of ET to PET is a useful measure of plant water supply in relation to plant water demand.It also can be used as an indicator of soil moisture conditions (22).It can be calculated as follows: (1) Where: ET represents EvapoTranspiration and PET is Potential EvapoTranspiration To calculate the value of DSI, standardization for values of the Ratio and NDVI, see table (2).
Where: NDVI is Normalized Difference Vegetation Index, represents temporal standard deviation, ̅̅̅̅̅̅̅̅ is Ratio average, and ̅̅̅̅̅̅̅̅ is Normalized Difference Vegetation Index average.
They have added Eq. 2 and Eq. 3, then standardized the result again to derive the DSI. (4) DSI combine NDVI and the ratio (ET/PET) for a single index extend from less than -1.5 (drier than normal) to greater than 1.5 (wetter than normal) (table 3) ( 16).̅ ̅ (6) Where P is the precipitation and ̅ is the average precipitation for a certain period.

Estimate Unknown Points
In this study, spatial interpolation techniques were used in Geographic Information System (GIS) package to cover the whole extent of country and extracting the zones.The ordinary kriging techniques was used to estimate values at unknown points by using known measurements and the continues surface data can be interpolated from the isolated point data such as weather station (3).After the extracted average total annual precipitation for each zone used to calculate the PPA using Eq.(6).

Evaluation the DSI Efficiency
The regression models were used between DSI and Percentage of Precipitation Anomaly (PPA), and the determination coefficient (R 2 ) was selected as a criterion for evaluating the efficiency of drought monitoring.Pearson correlation coefficient (R) in Equation ( 7) between DSI and drought indicators selected at study period was also calculated to Evaluation of drought patterns in Iraq and determining the most Susceptible areas (Zones) of this phenomenon.
Where:  4) and figure (2c) shows a significant fluctuation in precipitation amount from the average at all zones over Iraq, especially at Arid and Semi-Arid Zone when compared with Steppes and Desert zones, and this is clear from the large standard deviation value (STDEV) in this zone, this fluctuation is decreasing towards the Steppes and Desert zones result of to the decreasing precipitation in these zones during the study period.From time series of evapotranspiration, it can be seen decreasing in trend of evapotranspiration during study period, result decreasing in the trend of precipitation which leads increasing the capacity of the atmosphere to removal water from the surface through the processes of evaporation and transpiration assuming no control on water supply (PET).Followed Figure 4 Followed Figure 4 From Figure above it can be seen that the relationship between DSI-ET was positive while DSI-PET was negative, the reason as mentioned earlier that potential evapotranspiration (PET) is the capacity of the atmosphere to removal water from the surface through the processes of evaporation and transpiration assuming no control on water supply.While evapotranspiration is the quantity of water that is actually removed from a surface due to the processes of evaporation and transpiration and this is dependent upon the amount of water in surface as shown in section of evaluation of behavior of drought indicators and values of Ratio in table (2), which can be seen the minimum values of ratio were -0.03 and -0.01 in 2008 at Arid and Semi-Arid and Steppes Zones respectively while at desert zone was -0.01 in 2010 with minimum values of evapotranspiration and average precipitation see figures (2a) and (2c), which that mean the relationship between them are positive correlation while the values of potential evapotranspiration in 2008 and 2010 were maximum value during study period that mean the relationship between them are inverse correlation , see figure (2b).

Drought Events
The Drought events and most Susceptible Zone of drought in Iraq can be determined from the DSI time series and from precipitation anomalies, see figure (5) and table (4).The annual DSI and precipitation anomalies were extracted per Zone for the period 2000-2011 using same way used in sections (1.3.2) and (2.2).Every drought event is determined by unlimited negative values, see table (3).Figures (6) The Drought events also can be determined from precipitation anomalies.From table (4), it can be seen that there is a significant fluctuation in precipitation amount when compared the standard deviation of precipitation values for each zone with precipitation anomalies during the study period especially in seasons: The greater positive anomaly value was 215.5 mm at Arid and Semi-Arid Zone in (2002)(2003), which is greater than the standard deviation of precipitation at this zone (136mm), So, this is leads to the load the soil with water and increasing the vegetation cover, this is evident from the value of DSI in 2003 as a result of increased vegetation cover (NDVI) and thus increased evapotranspiration process (ET).The greater negative anomaly value was -232.9 mm also at Arid and Semi-Arid zone in (2007)(2008), which is greater than the standard deviation of precipitation at this zone.Also, from table ( 4 6) and table (4) we can conclude the Maximum and Minimum threshold of PPA for every Zone (the ratio of the standard deviation to the average precipitation for every zone), which that wet year if year exceed maximum threshold and dry year if that year exceed minimum threshold and between this maximum and minimum values are represent the normal fluctuation of precipitation for each zone.The threshold levels are 25% mm, 32% and 39% for Arid and Semi-Arid, Steppes, and Desert Zone for long-term study period respectively.Figure (8) be split into three climatic zones according to the rainfall factor, see figure (1): (i) Arid and Semi-Arid Zone where annual rainfall above 400 mm, Steppes Zone where annual precipitation of 200-400 mm, and Desert Zone where annual rainfall less than 200 mm ( represents meteorological parameters (drought indicators), y the DSI at different time scales, ̅ The mean of DSI during time scale, ̅ The mean of drought indicator during time scale and n is the number of samples RESULTS AND DISCUSSION Evaluation of the behavior of drought indicators: The Evapotranspiration (ET), Potential evapotranspiration (PET), and Seasonal Precipitation (PRE) data for the period (2000-2011) were aggregated into average total annual values except Precipitation was aggregated into average of total seasonal precipitation, divided into three climatic zones.Average evapotranspiration during the study period, increasing towards the Arid and Semi-Arid Zone due to the increase in the amount of Precipitation in this Zone compared to the amount of Precipitation in the Steppes and Desert Zones, as shown in the figures (2a) and (2c), while average Potential evapotranspiration increases towards the Steppes and Desert Zones due to lack of precipitation in these Zones and increase the net solar radiation at earth surface (Penman-Monteith equation) when compared with the Arid and Semi-Arid Zone, As shown in figures (2b),(2c).Table (
), The greater positive anomaly value at Steppes and Desert Zones were in (2002-2003) and (2005-2006) respectively, which is greater than the standard deviation of precipitation at these zones.So, it is clear that (2007-2008) is the drier season in all Zones, the wetter seasons are (2002-2003) at Arid and Semi-Arid and Steppes zone and (2005-2006) at Desert zone.Also, the Arid and Semi-Arid zone has greater fluctuation in precipitation amount from other zones during study period.Figure (7) shown Percentage of Precipitation Anomaly (PPA) during study period for three zones, it can be seen the maximum positive percentage value of PPA are 58.4% in 2000-2001 at Desert Zone, 40.4% and 35% in 2002-2003 at Arid and Semi-Arid Zones and Steppes respectively, and minimum negative percentage value of PPA are -63%, -60.5%, and -43.5% at Desert zone, Steppes and Arid and Semi-Arid Zones respectively in 2007-2008.From table (4), It can be seen the Desert zone is characterized by the greatest value of Precipitation from the average (Anomaly) in 2005-2006, but from figures (5), (6) and (7) noted That the greatest value at Desert zone was in (2000-2001), The reason for the amount of Precipitation during this season comparing with season 1999-2000, which lead to soil kept amount of water a result of the high amount of precipitation, therefore an increase of vegetation cover and increased evapotranspiration process during this season.So, the wetter year at Desert Zone was 2001, this is evident through the large value of DSI in 2001 and PPA in season (2000-2001).

Fig. 7 .
Fig. 7.The values of PPA for Iraq: Arid and Semi-Arid Zone, Steppes Zone and Desert Zone In order to study climate change at study area, Percentage of Precipitation Anomaly (PPA) was used as an index of drought monitoring and thus determining the most Susceptible areas of this phenomenon for long-term study period 1980-2015.From equation (6) and table(4) we can conclude the Maximum and Minimum threshold of PPA for every Zone (the ratio of the standard deviation to the average precipitation for every zone), which that wet year if year exceed maximum threshold and dry year if that year exceed minimum threshold and between this maximum and minimum values are represent the normal fluctuation of precipitation for each zone.The threshold levels are 25% mm, 32% and 39% for Arid and Semi-Arid, Steppes, and Desert Zone for long-term study period respectively.Figure(8) shows that the behavior of Percentage of Precipitation Anomaly is similar to his of previous figure for the first study period, with a difference in the average values for each zone of the study period.The maximum positive percentage value of PPA are 83.9% and 51.7% in season 1987-1988 at Steppes and Arid and Semi-Arid Zones respectively, while at Desert Zone was 81.5% in seasons 1994-1995 and 1997-1998.
Fig. 7.The values of PPA for Iraq: Arid and Semi-Arid Zone, Steppes Zone and Desert Zone In order to study climate change at study area, Percentage of Precipitation Anomaly (PPA) was used as an index of drought monitoring and thus determining the most Susceptible areas of this phenomenon for long-term study period 1980-2015.From equation (6) and table(4) we can conclude the Maximum and Minimum threshold of PPA for every Zone (the ratio of the standard deviation to the average precipitation for every zone), which that wet year if year exceed maximum threshold and dry year if that year exceed minimum threshold and between this maximum and minimum values are represent the normal fluctuation of precipitation for each zone.The threshold levels are 25% mm, 32% and 39% for Arid and Semi-Arid, Steppes, and Desert Zone for long-term study period respectively.Figure(8) shows that the behavior of Percentage of Precipitation Anomaly is similar to his of previous figure for the first study period, with a difference in the average values for each zone of the study period.The maximum positive percentage value of PPA are 83.9% and 51.7% in season 1987-1988 at Steppes and Arid and Semi-Arid Zones respectively, while at Desert Zone was 81.5% in seasons 1994-1995 and 1997-1998.

Fig. 9 .
Fig. 9. Annual DSI for Iraq: Arid and Semi-Arid Zone, (b) Steppes Zone and (c) Desert Zone The daily DSI was extracted for three stations to represent the zones: Sulaymaniyah represent the Arid and Semi-Arid Zone, Mousl represents steppes zone and Baghdad, represents the desert zone for the period (2000-2011), see figure (1).Figure (10) shows the time series of DSI in 2008 at three selected stations.The drought severity of DSI was -38.31 in Baghdad station (Desert Zone) and drought duration 35 time below threshold level with intensity -1.1.Also, the drought severity

Fig. 10 .
Fig. 9. Annual DSI for Iraq: Arid and Semi-Arid Zone, (b) Steppes Zone and (c) Desert Zone The daily DSI was extracted for three stations to represent the zones: Sulaymaniyah represent the Arid and Semi-Arid Zone, Mousl represents steppes zone and Baghdad, represents the desert zone for the period (2000-2011), see figure (1).Figure (10) shows the time series of DSI in 2008 at three selected stations.The drought severity of DSI was -38.31 in Baghdad station (Desert Zone) and drought duration 35 time below threshold level with intensity -1.1.Also, the drought severity