Regional Drought Monitoring in Zayandeh-rud Basin Based on Time Series Variations of the SPI and Satellite-Based VCI Indices

جغرافیا و مخاطرات محیطی

Article 1, Volume 6, Issue 4 - Serial Number 24, January 2018, Pages 1-22   

Document Type: Research Article

DOI: 10.22067/geo.v6i4.62601

Authors

Marziehsadat Mirahsani^{* 1}; Abdulrassul Salman Mahini¹; Alireza Soffianian²; Reza Moddares²; Reza Jafari²; Jahangir Mohammadi¹

¹Gorgan University of Agricultural Sciences and Natural Resources

²Isfahan University of Technology

Abstract

1. Introduction
Zayandeh-Rud Basin (ZRB) is one of the most strategic river basins of Iran where the majority of the basin has a typical arid and semi-arid climate with frequent droughts. Drought is a recurrent phenomenon that affects humid as well as arid regions to some degree which has large adverse consequences on the socio-economic condition of people living in drought-prone areas through its impact on water availability and quality and ecosystem health. Standardized precipitation index (SPI) is a station-based drought index which measures precipitation. The limitation of SPI is continuous spatial inability and coverage. On a regional scale, SPI cannot monitor and illustrate the detailed pattern of drought conditions spatially, particularly in regions with a high degree of spatial variability or inadequate meteorological stations. As the spatiotemporal identification of drought events is very complex, drought indices are mainly beneficial in order to monitor the impact of climate variability on vegetation (Khosravi & Akhund-Zadeh, 2015). Remotely sensed spectral indices can determine the quantification of vegetation drought. Moderate Resolution Imaging Spectro-radiometer (MODIS) has been used by various researchers from diverse disciplines globally. Monitoring vegetation using satellite imagery is also comprehensively applied to monitor and assess drought condition and etc. The Normalized Difference Vegetation Index (NDVI) (Tucker, 1979) is the oldest remotely sensed vegetation index in use and remains as the most widely used by the remote sensing community. The efficiency of NDVI in evaluation of vegetation response to drought has been tried out (Tucker & Choudhury, 1987) which provides a common measure of the health and condition of vegetation. Nevertheless, NDVI has some deficiencies; NDVI is not sensitive to soil color, atmospheric effects, and illumination and observation geometry. The limitation of NDVI in drought monitoring is that NDVI cannot indicate the drought severity alone because of obvious time lag of NDVI response to precipitation and little effect of weighty rainfall events later in the growing season (plant seed production period) on NDVI (Ganesh, 2007). Therefore, NDVI-Vegetation Condition Index (Kogan, 1995) normalizes NDVI for each pixel over time based on its minimum and maximum values. The VCI was applied successfully for drought monitoring by various researchers. Onset of drought and the duration, intensity and impact of drought can be measured and detected by the VCI. In addition, identifying the spatio-temporal variability of vegetation conditions associated with perturbation as a result of precipitation shortage is allowed by VCI. VCI has been compared with field reflectance measurements, biomass and vegetation density. It is proven that VCI could be a suitable sign of the impact of climate on health and conditions of vegetation. The clear benefit of VCI is simplifying of the computation because it does not need station observation data. As VCI is a satellite-based drought product, it can process globally near real-time data at a reasonably high spatial resolution. VCI is appropriate to make a relative assessment of variations in the NDVI signal by filtering out the contribution of local geographic resources to the spatial variability of NDVI. In addition, VCI evaluates variations in the NDVI signal during time through decreasing the effect of local ecosystem variables. This research aims to assess spatial distribution of VCI index through spatiotemporal monitoring of it using MODIS NDVI time series products in growth season over ZRB from April to October 2003 to 2014 (Mrahsani, Salman Mahini, Soffianian, Modarres, Jafari & Mohamadi, 2015).
2. Material and Methods
2.1. Study area
ZRB or Gavkhuni Basin is known as a very important river basin of Iran. ZRB with area of 41485.65 km2 (5.2% of the country's total area) is located between the 50° 24′ to 53° 24′ longitudes and 31° 11′ to 33° 42′ latitudes. Temperatures have been recorded to be hot in July, with an average of 30 ̊C, and with an average minimum temperature of 3 ̊C in January. Rainfall, which is generally very limited, is around 130 mm per year (Salemi, & Rust, 2004). In ZRB, the annual precipitation recorded ranges from 1500 mm (in snow which is only likely to be melted when temperature rises around April) in the western part to 50 mm in the eastern part (Sarhadi & Solatni, 2013).
2.2 Data Collection
2.2.1. Monthly MODIS/Terra NDVI
The Monthly MODIS NDVI time series product (MOD13A3, 1 km ×1 km) images of the study area were applied from 2003 to 2014.
2.2.2. Meteorological Drought Index
The SPI is a station-based index which calculates the probability of precipitation in different time scales. The SPI calculation for any location is based on the long-term precipitation record for a desired period. In this research, SPI values are calculated based on precipitation data collected from 26 meteorological stations around ZRB which had the same available beginning from January 2003 to December 2014.
2.3. Vegetation Condition Index (VCI)
VCI is calculated and normalized using long term NDVI ranging from 0 to 100. Low values indicate vegetation stress and median values indicate average condition while high value indicates the optimized and normal condition.
3. Results and Discussion
3.1. Metrological Drought Index (Monthly SPI)
The long-term SPI (12 months) were calculated based on precipitation data collected from 26 meteorological stations around ZRB which had the same available beginning from January 2003 to December 2014. The severity-duration and relative frequency of drought were then determined per time scale and station.
3.2. Vegetation Condition Index (VCI)
In the present study, the annual VCI time series maps were prepared in order to quantify drought from a long-term observation. It was found that sign and onset of drought can be clearly observed from the VCI maps from July to October 2003 and July to October 2005. The slope of fluctuation of the precipitation is slow during 2004 to 2006. From October to June 2007 the condition was improved all over the basin. Nevertheless, a severe drought condition prevailed all over growth season of the year 2008 all over ZRB. The slope of precipitation is slow during 2008 to 2009. The onset and extent of drought can be clearly observed from the VCI maps of April 2008 to end of growth season. Acute water stress is evident over the basin during June 2008 to April 2009. From April to May 2010, the condition was improved all over the basin. For a second time, a severe drought condition prevailed all over growth season of the year 2011 all over of ZRB. From April to May 2011, the condition was improved all over the basin, while in September the basin have faced drought again, this cycle is repeated in the end of growth season of the year 2014. Uneven distribution of rainfall causing spatial variation in vegetation health has a great influence on variation of agricultural yield of ZRB. Since the precipitations in the eastern, northern and central districts of the basin are much less than the western districts, the vegetation condition was visibly stressed in the Gav-Khuni wetland and agricultural lands in east of the basin. The vegetation health was only found normal in ZRB during the year 2007 and 2010. In 2008, the basin had been covered mostly by very high drought categories, followed by 2010 and 2014. High drought categories were mostly formed in 2011, followed by 2012 and 2013.In addition, in 2012, ZRB was mainly covered by relatively high drought, followed by 2014. The biggest area covered by moderately drought category belongs to 2012. The biggest area covered by normal category belongs to 2007, followed by 2006.
4. Conclusion
One of strengths of the MODIS sensor is the conciliation and reconciliation between reasonable spatial resolution and regular image acquisition. It is necessary to use remote sensing data beside other datasets from other sources because of the growing demand for data and escalation analysis. This causes an essential role for remote sensing techniques in the climatological and meteorological applications. The drought was assessed through monitoring vegetation condition using VCI; subsequently, its correlation with SPI was obtained to be 0.83, respectively. Then, the area of each category covered by VCI was calculated. Results indicate that VCI can present consistent results with strong correlation to SPI. In addition, results show usefulness of VCI in drought monitoring studies despite the wide range of climatic conditions in the region. In addition, the area of normal category covered by VCI in the basin is coincident during the time period except 2008 and 2011 during which severe meteorological drought had been occurred. However, it seems that vegetation condition is affected by facilitating irrigation of cultivated lands. In addition to extraction of groundwater and wells, the irrigation is influenced by opening the floodgate of Zayandeh-Rud Dam in the basin. Hence, it is essential to consider hydrologic drought in seasonal time scale to have a better image of water scarcity for different water demands such as agriculture and have a better water management and planning in future studies (Modarres & Silva, 2007).

Keywords

vegetation; Lag response; MODIS-NDVI; Pearson Correlation

References

باعقیده، محمد و دیگران؛ 1390. بررسی امکان استفاده از NDVI در تحلیل خشکسالی‌‌‌های استان اصفهان. مطالعات جغرافیایی مناطق خشک. 1(4)، 16-1.

خداقلی، مرتضی و دیگران؛ 1393. تحلیل روند گذشته و پیش‌بینی آینده خشک‌سالی در استان اصفهان. نشریه علوم آب و خاک-علوم و فنون کشاورزی و منابع طبیعی، دانشگاه صنعتی اصفهان. 18(67): 367-379.‏

خسروی، ایمان و دیگران؛ 1394. مدلسازی و پیش‌بینی سری زمانی شاخص‌های خشکسالی با روش‌های یادگیری ماشین به‌منظور مدیریت مخاطرات (مطالعۀ موردی: منطقۀ شرقی اصفهان). دانش مخاطرات، 2(1), 51-65.

سالمی، حمیدرضا و دیگران؛ 1383. سیمای کلی هیدرولوژی حوضه آبخیز زاینده‌رود. دو ماهنامه علمی-پژوهشی آب و فاضلاب. 15(2). 2-13.

رضایی بنفشه، مجید و دیگران؛ 1394. تحلیل خشکسالی کشاورزی استان آذربایجان شرقی با تاکید بر سنجش‌ازدور و شاخص وضعیت پوشش گیاهی. نشریه دانش آب و خاک. 25 (1). 113-123.

رضایی، پرویز و دیگران؛ 1393. تعیین روابط شدت، مدت، فراوانی و سطح درگیر با خشکسالی در شمال غرب ایران. نشریه جغرافیا و مخاطرات محیطی، (3) ۱۰، 111-127.

موذن زاده، روزبه و دیگران؛ 1391. پایش خشکسالی در کشت‌های غیرآبی با استفاده از تکنیک سنجش‌ازدور. مدیریت آب و آبیاری. 2(2). 39-52.

میراحسنی، مرضیه سادات و دیگران؛ 1395. تحلیل‌های خشکسالی و سری‌زمانی شاخص NDVI تصاویر ماهواره‌ای MODIS جهت پایش تخریب پوشش گیاهی در حوضه زاینده‌رود. دومین کنفرانس بین المللی اکولوژی سیمای سرزمین. اصفهان: دانشگاه صنعتی اصفهان. 5 آبان 1395. 11 صفحه.

http://www.civilica.com/Paper-IALE02-IALE02_105.html

میراحسنی، مرضیه سادات و دیگران؛ 1396. پایش مکانی- زمانی خشکسالی هواشناسی بر اساس پهنه‌بندی خوشه‌های ایستگاهی در حوزه آبخیز زاینده‌رود، نشریه مهندسی و مدیریت آبخیز، در نوبت چاپ.

میراحسنی، مرضیه سادات؛ 1396. تحلیل سری زمانی شناسایی روند تخریب سرزمین بر مبنای تبیین خصوصیات بیوفیزیکی مؤثر با تأکید بر پایش مکانی زمانی خشکسالی و تغییرات پوشش/کاربری اراضی. پایان نامه دکتری. استادان راهنما. عبدالرسول سلمان ماهینی و علیرضا سفیانیان. دانشگاه علوم کشاورزی و منابع طبیعی گرگان.

Ganesh, S. (2007). Investigation of the utility of the vegetation condition index (VCI) as an indicator of drought (Unpublished doctoral dissertation). Texas A&M University, the USA.‏

Kogan, F. N. (1995). Droughts of the late 1980s in the United States as derived from NOAA polar-orbiting satellite data. Bulletin of the American Meteorological Society, 76(5), 655-668.‏

Kogan, F.N., & Unganai, L. S. (1998). Drought monitoring and corn yield estimation in Southern Africa from AVHRR data. Remote Sensing of Environment, 63, 219-232.

McKee, T. B., Doesken, N. J., & Kleist, J. (1993). The relationship of drought frequency and duration to time scales. In Proceedings of the 8th Conference on Applied Climatology (Vol. 17, No. 22, pp. 179-183). Boston, MA: American Meteorological Society.

McKee, T. B., Doesken, N. J., & Kleist, J. (1993, January). The relationship of drought frequency and duration to time scales. In Proceedings of the 8th Conference on Applied Climatology (Vol. 17, No. 22, pp. 179-183). Boston, MA: American Meteorological Society.‏

Modarres, R., & da Silva, V. D. P. R. (2007). Rainfall trends in arid and semi-arid regions of Iran. Journal of Arid Environments, 70(2), 344-355.‏

Sarhadi, A., & Soltani, S. (2013). Determination of water requirements of the Gavkhuni wetland, Iran: A hydrological approach. Journal of arid environments, 98, 27-40.‏

Tucker, C. J., & Choudhury, B. J. (1987). Satellite remote sensing of drought conditionsm. Remote Sensing of Environment, 23, 243-251.‏

Tucker, C.J. (1979). Red and photographic infrared linear combinations for monitoring vegetation. Remote Sensing of Environment, 8, 127–150.