STUDY OF CLIMATE CHANGE AND ITS IMPACT ON A PART OF BRAHMAPUTRA RIVER BASIN

Abstract

Before any long-term planning and management of water resources, the assessment of climate change impacts at catchment level is essential to the living being and are of the first and foremost priority. Previous studies relevant to the topic has provided a broader view of potential impacts, but not efficacious enough to address temporal variability adequately nor useful in decision making, especially for Dikhow catchment which is a part of the large Brahmaputra river basin in North East India. This study was planned to quantify the effect of climate change on Dikhow catchment with the specific objectives to (1) downscaling of the GCM output for rainfall and temperature parameters; (2) trend analysis of past and future rainfall and temperature using GCM data; (3) study the impact of ENSO on precipitation and river flow for the Dikhow catchment; and (4) analyze the temporal variability of drought phenomenon in the Dikhow catchment in the backdrop of spatial vulnerability of ENSO effects. STATISTICAL DOWNSCALING Statistical downscaling consists of seeking statistical relationships between the variables simulated well by GCMs (predictors) and the surface climate variables (predictands). For estimation of future average monthly rainfall and temperature at selected stations in the study area, Multiple Linear Regression (MLR) method was employed. To downscale the predictands (precipitation and temperature), potential predictors were selected from the National Center for Environmental Prediction (NCEP) reanalysis dataset. Selected predictors displayed the best significant correlations, consistently with the precipitation and temperature over the study period (1961–2001). The daily climate data of A2 and B2 SRES scenarios of the HadCM3 general circulation model (GCM) were used for future projection of rainfall and temperature. For model development, about 70% of randomly selected data were used for training, and remaining 30% for testing of the MLR. Average values of Nash-Sutcliffe Efficiency (NSE) for training and testing of MLR for maximum temperature (Tmax) was 0.912 and 0.869, for minimum temperature (Tmin) 0.952 and 0.931, and for precipitation 0.829 and 0.811 respectively in the study region. The average correlation coefficient for training and testing for Tmax were 0.955 and 0.932, for Tmin 0.972 and 0.971 and for precipitation 0.913 and 0.913 respectively. The average values of root mean square error (RMSE) for training and testing period for maximum temperature ii were 0.034 oC and 0.170 oC respectively and minimum temperatures were Tmin 0.022 oC and 0.148 oC respectively. The average values of root mean square error (RMSE) for training and testing period for precipitation were 17.268 mm and 18.484 mm respectively. MLR performed well in downscaling of minimum temperature as compared to maximum temperature and precipitation. Future projection shows that precipitation, maximum and minimum temperature will increase in future at all stations for both A2 and B2 scenarios. The magnitude of increase in precipitation, maximum and minimum temperature will be higher for the A2 scenario than the B2 scenario. Among all seasons, winter minimum temperature having the highest rate of change, warming is more pronounced during night-time than day-time. Lower elevated place will be the hottest place in the future over the study area. TRENDS IN RAINFALL AND TEMPERATURE Investigation of trends and its magnitude in rainfall and temperature (past and future both) has been carried out using Mann Kendell (MK) test and Sen’s Slope Estimator (SSE). To identify the possibilities of any rational relationship between the magnitude of the trend and elevation over study region, the study area was divided into three physiographic zones, namely; Low Elevated Zone (LEZ), Moderate Elevated Zone (MEZ) and High Elevated Zones (HEZ), because it is expected that each zone will respond differently. Stations with elevations below than 100 m from m.s.l., falls under LEZ and also known as alluvial plain, stations between 101- 1000 m elevation from m.s.l., known as the MEZ and stations between 1001 to 2000 m elevations were classified as HEZ. Stations under different physiographic zones are Sibsagar (LEZ), Mon and Tirap (MEZ), and Mokokchung, Tuensang and Zunheboto (HEZ). Historically (from the year 1901-2002) annual precipitation is decreasing (highest rate was 1.846 mm/year) significantly (at 5% level of significance) for three stations (Sibsagar, Mon and Tirap) and monsoon precipitation is decreasing (highest was 1.787 mm/year) significantly for four stations (Sibsagar, Mon Tirap and Mokokchung) from the years 1901-2002.Year 1959 was identified as the breakpoint for the entire time series of 1901-2002, before that the trend was increasing and after that the trend was decreasing for annual and monsoon precipitation time series. Further during entire time series (1901-2002), significant negative correlation was obtained between precipitation (annual and monsoon) trend magnitude and elevation of the study region. Significantly positive correlation between iii elevation of the study area and monsoon precipitation trend magnitude for the earlier period 1901-1959 and significant negative correlation between elevation of the study area and monsoon precipitation trend magnitude for later period 1960-2002 were identified. Further significantly rise for future precipitation were observed over the region for both A2 and B2 SRES scenarios, winter precipitation is likely to increase more (59.18 to 69.44 % under A2 and 47.71 to 54.90 % under B2 over the catchment) as compared to summer and monsoon precipitation. Significant rise in historical (1901-2002) maximum and minimum temperature, time series were observed on both annual and seasonal basis for all stations in the study area. Maximum temperature having higher rate of rising is observed than the minimum temperature for annual, summer and monsoon time series, while in winter season minimum temperature shows higher rate of rising rather than the maximum temperature (1.02 oC over 102 years). For Diurnal Temperature Range (DTR), monsoon season shows significant trend for four stations of the study area, magnitude of significant rising trend varies from 0.0016 oC/year to 0.0021 oC/year. Extreme values of annual Tmax and annual Tmin were also considered for this study, in which maximum and minimum values of annual Tmax are denoted by TXx and TXn respectively. Similarly, maximum and minimum values of Tmin are denoted by TNx and TNn respectively. Extreme temperature indices (TXx, TXn, TNx and TNn) also show the significant warming over the catchment. TXx shows the greatest trend over LEZ (1.1oC during 102 years), followed by MEZ (1oC during 102 years) and HEZ (0.91 oC during 102 years). In the same manner rate of trend of TXn is highest over LEZ (0.5 oC during 102 years) followed by MEZ (0.4 oC during 102 years) and no significant trend identified over HEZ. TNx shows the greatest magnitude over LEZ (0.5oC during 102 years) followed by MEZ (0.4 oC during 102 years) and least over HEZ (0.34oC during 102 years). TNn also shows a strong significant trend over LEZ (0.7oC during 102 years) followed by MEZ (0.6 oC during 102 years) and no significant trend over HEZ. The trend magnitude of extreme temperature indices and elevation shows a significant relationship while the trend magnitude of average Tmax, Tmin and DTR, does not show the noticeable connection between warming rate and elevation of the study area. Further, future Tmax and Tmin is increasing sharply for both A2 and B2 scenarios. A2 scenario shows higher rate of rise than the B2 scenario. Tmax is likely to increase by 0.038 oC/year, 0.036 oC/year, 0.035 oC/year and 0.01oC/year for winter, summer, annual and monsoon season respectively under A2 scenario. Minimum iv temperatures are projected to increase by 0.053 oC/year, 0.051 oC/year, and 0.049 oC/year for winter, summer and monsoon respectively for A2 scenario. DROUGHT ASSESSMENT Droughts are periodic climatic events and are recognized as a major limiting factor for the regional economic development by affecting agriculture, water resources and food production. Precipitation departure from 1950-2002 shows that an annual precipitation deficiency varied from 21% to 27%, while monsoon precipitation deficiency varies from 25% to 30% and the average frequency of drought occurrences is once in every 13 to 18 years in both annual and monsoon precipitation in the study area. Further, dry and wet years were identified using the Standardized Precipitation Index (SPI) at 3, 6 and 12 months time scales for the period of 1950-2002 for June, July, August and September months. Years 1956, 1978, 1979, 1980 and 1996 at SPI 3, years 1967, 1970, 1979 and 1980 at SPI 6, and year 1979 at SPI 12 were observed as extreme drought years from 1952 – 2002 in the study area. No extremely wet events were found at SPI 3 and SPI 6 months time scale, while at the SPI 12 month time scale year 1990 was found to be an extremely wet year at Sibsagar and Tirap stations from 1952 – 2002 during June, July, August and September months. Moreover, assessment of drought severity on crop production, shows that the crop production was less during year 2000 (severe drought year) than year 1999 (normal year). The yield of kharif crops like rice, maize, jwar, and bajra is less during the year 2000 and vis-à-vis during 1999. Its effect can also be observed over the Rabi crops like rapeseed and linseed. For rice crop, most drought affected stations were ranked as Tirap, Zunheboto, Tuensang and Mon. ENSO IMPACTS ENSO and Precipitation Correlation analysis between ENSO and precipitation indicated a significant negative correlation in monsoon season for three stations (Sibsagar, Mon and Tirap) in the catchment. Therefore, further analysis was carried out for monsoon season precipitation only. The CDF of standardized monsoon precipitation during El Niño, La Niña, and full time series shows that the exceedance probability for average monsoon precipitation is less in El Niño years and more in La Niña years. v It was also observed that monsoon precipitation of alluvial plains (LEZ) and moderate elevation zone (MEZ) were affected by ENSO in the catchment. Over high elevation zone no significant impact of ENSO was observed. ENSO and River flow Correlation analysis between ENSO and river flow (1988-2008) indicated non-significant negative correlation in monsoon river flow at the outlet (Sibsagar) of the catchment. ENSO and Drought The years 1950-2002 were categorized into three types of El Niño years on the basis of the intensity of SST anomalies. During the study period, 12 strong, 13 moderate and 9 weak El Niño years were identified. El Niño has a history of adversely impacting rainfall in India during monsoon season and cultivation in the selected regions is rainfed in most of the area. For this purpose monsoon season is only considered in the analysis of El Niño years. For Teleconnections between El Niño and drought, Standardized Precipitation Index (SPI) was used to explore the temporal and spatial variability of drought in the catchment. Droughts were analyzed mainly at 3, 6 and 12 months time scale for agriculture and hydrological purpose and classified into three categories, extreme, severe and moderate drought on the basis of SPI values. Maximum drought events were identified in SPI 3 months time scale, in which, year 1956 (Sep), 1978 (Sep), 1979 (June, July, Sep) and 1996 (Aug) were observed as extreme drought years in the study area out of which only the years 1956 and 1978 were categorized as weak El Niño year. The catchment also experienced smaller intensity drought in the years 1959, 1962, 1967, 1971, 1973, 1980, 1984, 1989 and 2000 and falls under the category of moderate to severe drought in the study area, out of which the years 1973 and 1959 were noticed as only El Niño (strong) years. Drought events were identified more in lower (Sibsagar) to moderate elevation zone (Mon and Tirap) in the complex catchment topography at three months time scale. Moreover, the year 1979 which was categorized as a non - El Niño year, observed as an extreme drought year in all stations and at all time scales in the selected study area (average annual rainfall was 1643 mm in year 1979). However, in previous studies, it was noticed that the year 1997 was the Century’s strongest El Niño year, but it did not cause a drought in whole India as well as in the Dikhow Catchment. No strong observation was found in relationship with El Niño intensity and drought severity. vi Further, the year 1990 was the extremely wet year identified at SPI 12 months time scale which was not a La Niña year. Additionally, Results show that not all wet years were La Niña years out of which at SPI 3, years 1954, 1955, 1964, 1973, 1974, 1988, 1993, 1998 and at SPI 6, years 1954, 1955, 1964, 1973, 1974, 1988, and 1998 and at SPI 12 years 1955, 1956, 1974 and 1989 were observed as wet years as well as La Niña years. The study shows that few drought years were linked with El Niño years. Relatively, number of wet years were more which are associated with La Niña years in the study period over the study area, hence it can be concluded that the region is impacted more by La Niña than El Niño. However, it should also be noted that large deficits/excess do occur in the absence of El Niño / La Niña. The present study will afford vision to policymakers, scientists and stakeholders for planning, mitigation and adaptation measures to counter the adverse impacts of climate change. It will also be helpful to provide rational information to agricultural and hydrologist for proper planning and scheduling of crops and water management under projected climate change regime.