Drought Assessment in Tel Watershed: An Integrated Approach Using Run Analysis and SPI

S. Sangita Mishra and R. Nagarajan
Centre of studies in resources engineering,
Indian Institute of Technology. Bombay
Powai, Mumbai, India

e-mail: sangita_csre@iitb.ac.in

Abstract — Drought can be a disastrous natural phenomenon that can have significant impact on socio-economic, agricultural and environmental spheres. Since droughts are natural phenomena, their occurrence cannot be predicted with certainty and thus it must be treated as a random variable. Once drought duration and magnitude have been found objectively, it is possible to plan for the transport of water in known quantities to drought-stricken areas, either from alternative water resources or from water stored during wet periods. For this study, the summation of deficits over a particular period is referred to as drought magnitude. Drought intensity is the ratio of drought magnitude to its duration. These drought properties at different truncation levels provide significant hydrological and hydro-meteorological design quantities. In this study, the run analysis and Standardized Precipitation Index (SPI) were used to investigate drought behaviors in Tel watershed, which is about 2,756 square kilometers and lies between 19° 17′ and 20° 00′ N latitude and 82° 30′ and 82° 59′E longitude in Kalahandi district of Odisha, India. Drought duration (dry spells), magnitude and severity, along with the SPI values, are presented to depict the relationships between drought duration and magnitude.

I. Introduction

Droughts are hydro-meteorological events affecting vast regions and causing significant structural and non-structural damages. Drought predictions may prevent these adverse consequences to a significant extent. In order to reach such a target, it is necessary to develop a method of monitoring based on the available past experiences as well as on environmental conditions.

Figure 1: Location map of the Tel watershed, Odisha

Figure 1: Location map of the Tel watershed, Odisha

There are several indices that measure to what extent precipitation for a given period has deviated from historically established norms. Although none of the major indices is inherently superior to the rest in all circumstances, some indices are better suited than others for certain uses. The Standardized Precipitation Index (SPI) was designed to quantify the precipitation deficit for multiple time scales, which reflect the impact of drought on the availability of different water resources. Soil moisture conditions respond to precipitation anomalies on a relatively short time scale, while groundwater, stream flow, and reservoir storage reflect the longer-term precipitation anomalies as in [12]. For these reasons, McKee et al. [2] originally calculated the SPI for 3-, 6-, 12-, 24-, and 48-month moving average time scales. The SPI is probability-based, as in [4], and was designed to be a spatially invariant indicator of drought ([7], [8], [9]) which recognizes the importance of time scales in the analysis of water availability and water use [3].

Run analysis is proved to be the simplest methodology along with SPI to provide an objective identification and characterization of drought events as in [11].

The main purpose of this study is to identify various drought properties on the basis of run analysis and SPI with applications to five stations (Bhawanipatna, Dharangarh, Junagarh, koksara and Jayapatna) in Tel watershed of Odisha, India. Empirical relationships are provided through scatter diagrams between the drought magnitude and length.

II. Study Area

The study area is one of the watersheds of Tel river located in the Kalahandi district of Odisha, which is one of the worst drought-affected districts of India (Fig.1). The watershed covers an area of 2,756 square kilometers and lies between 19° 17′ and 20° 00′ N latitude and 82° 30′ and 82° 59′E longitude near Bhawanipatna region of Kalahandi district. The study area experiences a tropical wet and dry climate. The wet season (June–September) is much shorter and receives low precipitation from the south-west monsoon. The remaining months of the year are generally dry, because the area does not receive any precipitation from the north-east monsoon.

TABLE 1: SPI Classes

TABLE 1: SPI Classes

III. Methodology

In this study, the monthly rainfall data of 5 meteorological stations over a period from 1965-2008, which are well distributed in the Tel watershed area, were used for drought analysis. The identification and assessment of drought severity were done using the SPI. The SPI is calculated from the monthly precipitation record by first fitting the gamma probability distribution function and then transforming that into a normal distribution so the mean SPI is set to zero ([2], [5]). The Table 1 shows the SPI classes. In this study, the SPI is calculated for 1-,2-,3-,6- and 9-month time scales.

Figure 2: Drought characteristics using the run theory for a given Threshold level

Figure 2: Drought characteristics using the run theory for a given threshold level

Temporal drought characters (i.e duration, severity and intensity) were determined using the run analysis as in [13]. The run theory has been applied in several drought models and analyzes ([1], [6], [10], [14], [15]). A run is defined as a portion of time series of drought variable Xt, in which all values are either below or above the selected truncation or threshold level of X0; accordingly it is called either a positive run or a negative run (Fig. 2). In this study, the threshold level was obtained by dividing the average annual precipitation in the study area by the mean number of rainy days in that area. The use of absolute threshold allows an easier comparison of the results from different climatic regions. A period along the time series is classified as dry if more than 7 consecutive dry days occurred. The drought duration is considered as the time period between the initiation and termination of a drought. Drought severity/magnitude is calculated as the cumulative deficiency of precipitation below the threshold level and drought intensity is obtained by dividing the drought severity/magnitude by the drought duration.

Figure 3: 1 month SPI values for Junagarh station

Figure 3: 1 month SPI values for Junagarh station

IV. Results and Discussion

SPI indices were determined using monthly total precipitation series at the five meteorological stations and results are discussed for May-October months. Since negligible rainfall occurred during November to April months, results of these months are not discussed here. The analysis showed that in 1-month scale, extreme drought was experienced in Junagarh with an SPI value of -2.5 (Fig. 3) in the month of July 2007. July was the most critical month, experiencing maximum number of extreme and severe droughts in the 5 meteorological stations. May followed July in experiencing severe drought and moderate drought in 9.78% and 13.63% of the total 44 years, respectively. Drought magnitude was increased with 2-month scale as compared to the 1-month scale. The highest magnitude observed was -2.74 in the month of July in Dharamgarh during 2002. Analysis showed that 2002 rainy season was affected by severe drought with all the months having negative SPI value (Fig. 4).

Figure 4: 2 month SPI values at all stations in the year 2002

Figure 4: 2 month SPI values at all stations in the year 2002

The lowest extreme drought event was experienced during May, with the value of -2.02 in Junagarh in a 3-month time scale. May was found to be the most critical month, with 43.15% of severe drought and 25% of moderate drought of the total area, while October was the least critical month with 25% of moderate drought, 11.36% severe drought and only 2.27% extreme drought during the investigation period. The maximum number of observed drought events occurred in May, followed by July and August.

August was the most critical month, considering both the number of dry years and the extending area of extreme drought events in the 6-month category. Ten years were successively dry during May at the Jayapatna station from 1999 to 2008 (Fig 5). By the end of June 2002, extreme drought covered 78.5% of the total area and the remaining areas stayed under severe drought conditions with a negligible moderate drought. The drought was experienced with the highest SPI value of -2.92 in Dharamgarh in July 2002. In the 2002 rainy seasons, another dramatic prolonged event happened; all study areas experienced drought conditions for all investigated months. During this period, the drought event started with an SPI value of -1.05 as the highest in May, reached peak value (-2.92) in July at the Dharamgarh station and ended in August with a value of -0.95. Analysis showed that the years 1966, 1972, 1979, 1987 and 2002 were the most drought-affected years by droughts during the investigation period and thus affected agriculture production in the entire study area.

Figure 5: 6 month SPI values at Jayapatna station

Figure 5: 6 month SPI values at Jayapatna station

In the 9-month time scale, every drought category was observed in the study area. October was found to be the most vulnerable month to drought, where 55.45% of years were affected by drought during the investigation period. Similar to the sixth month droughts, the year 2002 was found to be the driest year with an SPI value of -3.05 in July in the Dharamgarh station (Fig. 6). The SPI values were negative for all the analyzed months.

The results of SPI analysis indicated the increasing trend of drought severity from the year 1999 to 2008 in all the five stations ( Fig. 3-6); hence a further study was attempted to find out the magnitude, duration, intensity of drought and the number of dry spells for this period in each station of Tel watershed using run analysis.

Figure 6: 9 month SPI values at Dharamgarh station

Figure 6: 9 month SPI values at Dharamgarh station

The maximum number of dry spells was 52 days in Junagarh station during the period under observation. The drought intensity was highest in Bhawanipatna station (i.e 20.42) in the year 2000, with the highest drought magnitude of 630.5. Bhawanipatna and Junagarh stations also recorded the highest drought duration of 52 days in the years 2002 and 2008, respectively.

The relationship between the magnitude and duration of droughts is established and empirical relationships are provided using scattered diagrams as shown in Fig. 7. These drought features refer to past observations, but their statistical parameters are useful for what-if analysis, conditionally valid also for the future. For instance, one can find drought magnitude, which corresponds to a given drought duration from the drought magnitude-duration graphs. This corresponds to the average water need in a critical period that should be met using “external” water resources, i.e accumulated in other time periods or spatial locations (e.g water stored in a reservoir during a wet period, or water transfer from another catchment).

V. Conclusion

Figure 7a-7e:  Magnitude-duration curve for the meteorological  stations of  Tel watershed.

Figure 7a-7e: Magnitude-duration curve for the meteorological stations of Tel watershed.

This study was focused on presenting a framework of methodologies for the assessment of drought occurrences as well as the identification of various drought characteristics, such as magnitude, duration and intensity at the 5 meteorological stations of Tel watershed, Odisha, India. The SPI computed at various time scales was used as an indicator of drought severity. The SPI analysis indicated that the years 1966, 1972, 1979, 1987 and 2002 were the years most affected by drought during the investigation period and thus affected agriculture production in the entire study area. The year 2002 was found to be the driest year. The highest SPI value for the 9 month drought ended in May and caused the lowest cereal yield recorded. Run analysis was carried out using the daily rainfall data for the period 1999-2008 and the relationships between drought duration and magnitude are provided in the form of scatter diagrams with the best straight line fits. The drought magnitude can be obtained corresponding to a given drought duration from the magnitude-duration graphs, which gives an idea about the water need in a critical period in a watershed for different purposes. These inferences will be helpful for better water management and agricultural practices in the similar areas around the world.

References

[1] A.K. Mishra, V.R. Desai, V.P. Singh, “Drought forecasting using a hybrid stochastic and neural network model,” J. Hydrologic Eng. ASCE, vol. 12 (6), pp. 626–638. Dec. 2007.

[2] B. McKee, N.J. Doesken and N. Kleist, “The relationship of drought frequency and duration to time scales.” In proc. of 8th Applied Meteorology,1993, p. 179-184.

[3] B. McKee, N.J. Doesken and N. Kleist, “Drought monitoring with multiple time scales,” In proc. Ninth Conference on Applied Climatology, 1994, p. 233–236.

[4] C.H. Chung and J.D. Salas, “Drought occurrence probabilities and risks of dependent hydrologic processes,” ASCE Journal of Hydrologic Engineering, vol 5, pp. 259-268, July 2000.

[5] D. C. Edwards and T.B. McKee, “Characteristics of 20th century drought in the United States at multiple time scales,” Colorado State University, Dept. of Atmospheric Science, Fort Collins, Colorado, Climatology Rep 97-2., pp. 155, 1997.

[6] H.A. Loaiciga, R.B. Leipnik, “Stochastic renewal model of low-flow streamflow sequences,” Stochastic. Hydrology and. Hydraylics, vol.10 (1), pp. 65–85. 1996

[7] I. Bordi, S. Frigio, P. Parenti, A. Speranza and A. Sutera, “ The analysis of the Standardized Precipitation Index in the Mediterranean area: regional patterns,” Annali di Geofisica, vol.44, pp. 979–993, Dec. 2001.

[8]I. Bordi, K. Fraedrich, J.M. Jiang and A. Sutera, “Spatio-temporal variability of dry and wet periods in eastern China,” Theoretical and applied climatology, vol. 79, pp. 81-91, Sept. 2004.

[9] I. Livada, and V.D. Assimakopoulos, “Spatial and temporal analysis of drought in Greece using the Standardized Precipitation Index (SPI),” Theoretical and applied Climatology, vol. 89, pp. 143-153. Dec. 2007.

[10] J.A. Dracup, K.S. Lee, E.G. Paulson, “On the statistical characteristics of drought events,” Water Resources. Reserch, vol. 16 (2), pp. 289–296. Oct. 1980.

[11] S. Sirdas and Z. Sen , “ Spatio-temporal drought analysis in the Trakya region, Turkey,” Hydrological Sciences Journal, Vol. 48(5) October 2003.

[12] T. Tonkaz, “Spatio-temporal assessment of historical droughts using SPI with GIS in GAP region, Turkey,” Journal of applied sciences, vol. 6, pp. 2565-2571, 2006.

[13] V Yevjevich, “An objective approach to definition and investigation of continental hydrologic droughts,” Colorado State Univ., Fort Collins, Colorado, USA. Hydrology Paper 23, 1967.

[14] Z. Sen, “Wet and dry periods for annual flow series,” J. Hydraulic Eng. Div, ASCE vol. 102, pp. 1503–1514. 1976.

[15] Z. Sen, “Statistical analysis of hydrologic critical droughts,” J. Hydraulics Div., ASCE vol. 106 (1), pp. 99–115. 1980.

Authors

S. Sangita Mishra, Research scholar in CSRE, Indian Institute of Technology, Bombay and working on droughts and natural hazards assessment and mitigation in the chronically drought affected regions of India.

R. Nagarajan, Associate Prof. in Center of studies in resources engineering in Indian Institute of Technology, Bombay and has been working on the natural hazards assessment and mitigation in the chronically drought affected regions of India since last 8-10 years. His work has been published in many journals. He is also the author of three books on Drought assessment and landslides and avalanches.

Topic:

Tags: