ASSESSMENT OF SOIL, SEDIMENT AND WATER QUALITY IN AGMCULTURAL MICRO WATERSHEDS

Abstract

Chemical pollution generated by agricultural, industrial and municipal activities has contaminated soil, ground water and surface water system worldwide. Amongst these activities, agriculture is the single largest user of fresh waterresources, using a global average of 70% of all surface supplies, and is a major cause of degradation of surface and ground water resources through erosion and chemical run off. Rainwater composition plays an important role in scavenging soluble components from the atmosphere and helps us to understand the relative contribution of different sources of atmospheric pollutants. Subsequent to a storm event, rapidly moving runoff creates flooding and transports high levels of sediment bound pollutants and dissolved contaminants into surface water. Increased runoff also causes erosion, resulting in the degradation of aquatic habitats and the accelerated deposition of sediments into rivers and reservoirs The sequence characterizing change in surface water quality is displayed by an increased ion concentration, followed by an increased algae growth, reduced water clarity, water treatment problem, reduced Oxygen in the water, altered fisheries, fish kills and toxins from Cyanobacteria affecting human beings and animal health. Although erosion has occurred through the history of agriculture, it has intensified in recent years. Soil erosion due to water has been observed as the main cause of land degradation in Doon valley, which receives about 1240 mm rainfall in monsoon season through high intensity storms. The present work aims to study the soil, sediment and runoff water quality generated from experimental agricultural micro watersheds in outer Himalayas at Central Soil &Water Conservation Research &Training Institute Research farm which is located in Selakui village in Dehradun District of Uttaranchal State, India. The farm is situated within 30° 20" Nlatitude and 77° 5T Elongitude on National Highway No. 72, 20 km from Dehradun. Six gauged experimental micro watersheds were selected at the experimental research farm of Central Soil and Water Conservations Research &Training Institute, Dehradun employing following runoff control as two watersheds having 2% land slope (control without runoff treatment, vegetative barrier treatment) and four watersheds having 8% land slope (Graded bund treatment, vegetative barrier and graded bund treatment, vegetative barrier treatment and control without treatment). The soil samples, collected from the watersheds during pre and post monsoon seasons and analyzed for soil texture, pH, E.C, major inorganic ions, nutrients, trace elements, bioavailability of trace elements and total organic carbon. The analysis showed that the soil texture for 2% soil slope watersheds was loam type and for 8%, it was sandy loam type. This gradual shift in the nature of soil texture could be attributed to the loss of fine particulate matter from higher land slope watersheds. Soil pH values varied from 6.50 to 6.30 while soil E.C. varied in the range of 313-199 uS/cm , 2% slope watersheds exhibiting higher E.C. values. This could be directly related to the extent of mineralization in these watersheds and the varying rates of soil erosion under different slope and treatment conditions. The concentration ofions followed the trend, K+I> Ca+2 >Mg+ >Na+ and HC03" > CI"1 >S04'2. Considering the normal level in the agricultural soil, the studied soils displayed low nitrogen levels which reflected the ability of nitrogen compounds to escape through different paths like runoff, subsurface leaching and plant consumption. On the other hand, phosphorus levels were observed to be higher which could be attributed to factors like settling (specially under the availability of calcium) as well as low solubility of phosphorus complexes. It was also found that agood correlation existed amongst K+1, HCO3"1 and SO4"2 ions which indicated the role of multiple sources in establishing the present constitution of soil. The order of occurrence of trace elements was Mn >Fe >Zn >Cu >Pb >Cr >Cd, which related to the characteristic soil formation. In addition, adifference of several orders of concentration was observed between Mn, the most abundant and Cd, the least abundant metal. However, based on comparison of this study with other soils worldwide, the soils of the study area appeared to be in the non-polluted range. Strong correlation among theses heavy metals (mostly larger than 0.70) indicated common geogenic sources. The order of the bioavailability of trace elements was Mn >Cu >Pb >Fe > Zn > Cd > Cr. The TOC levels in 2% slope watersheds (both control and treated) were observed to be higher than in 8% slope watersheds. Further, treated watersheds also exhibited higher TOC levels than control watersheds. Among the cations, post-monsoon reduction in K+1 was observed to be the least in both the 2% slope and 8% slope watersheds this clearly exhibited typical geochemical properties of the soil showing adequately enrichment with K+1 as indicated earlier. Na+1, on the other hand, was observed to get reduced substantially, specially in 2% slope watersheds. Retention of Na+1 appeared low as the general composition also indicated very low Na values. Among anions, least reduction in HCO3'1 , on one hand could be attributed to geogenic reasons as well as to external inputs during rainfall. Whereas regular application of P- fertilizers alongwith the low solubility of Pcompounds apparently resulting in offsetting of the losses in P04"3. Higher reduction in N03"', on the other hand, could be attributed to high level of its solubility in water. Among the trace elements, 8% slope watersheds exhibited higher post-monsoon reduction as compared to 2% slope watersheds, which reflected the influence of degree of slope. Increasing scale oftreatment appeared to result in lesser reduction oftrace elements. The loss of Pb was the largest as compared to other trace elements. The 2% slope watersheds also exhibited a higher reduction in TOC as compared to the 8% slope watershed. The treatment "Graded bund with Vegetative barrier" emerged as most effective means to check TOC loss. Calculation of "Enrichment Factor" for inorganic ions and trace elements considering Fe as reference element) indicated a loss in case of some elements (EF <1), possibly through percolation or by runoff. Anthropogenic impacts were apparent in case of Zn (addition of fertilizers like ZnS04 etc.) and Cu mostly, and Mn for some watersheds (on the basis of estimation of Enrichment Factor considering constituents of earth's crust and Cd as reference element). Rainwater samples, collected for a number of selected events from the study area, showed a descending order of concentration of ions as Ca+2 > Mg+2 > Na+1 > K+1 and CI'1 > HCO3"1 > NO3" > SO4' . Calcium was observed to be the dominant cation while CI"1 emerged as the dominant anion in the rainwater samples. High correlation among various constituents indicated that these ions may come from the same source. Computation of the "Neutralizing factor" for rainwater indicated high buffering capacity and very low possibility of acid rain in this area. The calculation of sea and non-sea fractions showed that the rainwater was not affected by sea sources justifying the fact that the study site was very far away from the seashore. The HCO3"1 load was the highest and main component of the total load from the rainwater. This also reflected to the nature oflocal dust suspended in the atmosphere. 111 The analysis of storm water runoff showed that there were no significant differences in pH value (6.5 - 6.7). The Median E.C. values for 2% slope watersheds were observed to be greater than that for 8% slope watersheds. The highest median E.C values was observed in 2% slope watersheds having vegetative barrier treatment and the lowest in 8% slope control watershed this could be related to the ionic composition of the parent soil as well as the solubility of ions derived from the soil of different watersheds The shape of radial plots of median ions concentration showed that the chemical species of the runoff water were similar in both 2% slope watershed having vegetative barrier treatment and 8% slope watershed having vegetative barrier and graded bund treatment (calcium chloride type of ionic composition). These watersheds also had the largest radial plot size indicating the largest level of chemical composition as well as a high tendency of losing these ions. The watersheds displaying the highest runoff level, viz. MWS3 and MWS6, showed similar shape of radial plots (Ca+2 and CI'1 concentration almost similar to concentration of Mg+2 and NO3'1). Runoff from 2% slope watersheds (both control and vegetative barrier type) exhibited higher Ca+" concentration in comparison to the 8% slope watersheds, whereas Mg+2 concentration exhibited an opposite trend. It was also found that the 2% slope watersheds yielded higher concentration of Potassium than 8% slope watersheds. This reflected the differences in soil ion composition of both type of watersheds. The median sodium concentration were low in all the watersheds when compared to the concentration values of other cation, apparently because of higher sodium activity and solubility. The chloride ion was observed as the dominant anion in runoff water in all the watersheds. Median Bicarbonate (HCO3"1) concentration values in the runoff water from the 2% slope watersheds were found higher than those from 8% slope watersheds. This related to the initial concentration of this ion in the parent soils. The sulfate (S04"2) concentration in runoff water was generally less than Chloride and Bicarbonate concentration and it also showed low variability among the various watersheds. However, its concentration for runoff from 8% slope watersheds was found slightly higher than 2% slope watersheds. Total Kjeldahl Nitrogen (TKN) concentration ranged from 105 to 136 mg4, the median values not showing high variability amongst different watersheds. The nitrate concentration showed a similar behavior like that of TKN in all the watersheds. Nitrate concentration ranged from 2.03 to 35.40 mg/1.,. The median nitrate concentration in runoff water varied from 7.32 to 8.74, and the maximum was observed in control treatment watershed of 2% IV slope. The difference in the concentration of nitrogen compounds was apparently affected by many factors like, the oxidation, uptake by plants and organisms and interaction of fertilizers with soil solution.. The concentration ofdissolved phosphate (P-P04) in runoff water varied from 0.998 to 0.401 mg/1. The difference in concentration directly reflected the availability of water as solvent, due to which watersheds with high runoff exhibited high concentration of P04"3. Dissolved organic carbon concentration in runoff ranged from 12.30 to 15.88 mg/1. Median concentration of dissolved organic carbon was observed to be highest in the control watershed of 8% slope. Generally occurring low values of dissolved organic carbon from different watersheds could be related to almost no use of manure based fertilizers in these sites and loss of top. soil. Application of analysis of variance (ANOVA) on the data indicated significant differences (P< 0.05) among Event mean concentration (EMC) values of E.C for various runoff treatments in 8% slope watersheds. Further, the analysis showed that the 2% slope watersheds exhibited higher event mean concentration values of E.C in comparison to the 8% slope. This reflected fast washoff of ions in watersheds displaying lower slope and scale of treatment resulting in adecrease in the time of dissolving ions from these watersheds. Further, significant differences ( P< 0.05) in EMC values of E.C was observed between years 2002 and 2003 which could be related to varying rainfall intensity and raindrop impact on the top soil. Significant difference in EMC values of TDS, Ca+2, K+1, CI'1, HCO/1 and N03'' was observed (P< 0.05) while non-significant difference was observed in case of EMC values for Na+1, Mg+2 and S04'2. In general, the higher scale of runoff treatment exhibited the higher values of EMC excluding P04-\ The 2% slope watersheds exhibited generally higher EMC values than for 8% slope. The EMC values were significantly different (P< 0.05) between two study years for all constituents excluding P04 and HC03, for whom the values in the year 2003 were higher. Only three trace elements viz. Zn, Fe and Pb appeared in the runoff water. Like in case of major ions, the 2% slope watersheds exhibited larger EMC values of Zn, Fe and Pb than those in 8% slope watersheds. The EMC values of Zn in runoff water were apparently affected by runoff control treatments, as significant difference emerged among watersheds (P< 0.05), the low scale runoff treatments displaying the larger values. This could be attributed to the large quantity of runoff water generated in low scale treatment. Asignificant difference was noticed in EMC values of Zn, Fe and Pb between the two years. Cluster analysis was attempted for each watershed to classify the generated information on all the physicochemical parameters into similar cluster classes. Unique clusters emerged for all watersheds constituents; pH value and TOC generally appeared in the same cluster while base inorganic ions have greater distance from pH value, which indicated the role of TOC in decreasing PH values. The Kand CI ions were in same cluster for all watersheds, which related to similaractivity of these ions in solutions. Export coefficients and Loading rates were calculated for all watersheds The 2% slope watershed having vegetative treatment exhibited the highest export coefficients in comparison to all cations excluding Na. The 2% slope watershed having vegetative treatment and the 8% slope watershed having vegetative and graded bund treatment type exhibited the highest loading rates. Power equation was used to estimate the dissolve loads generated from watersheds for selected parameters. While attempting regression between loads and runoff volume, Good relationship was found between runoff volume and loads of TDS, N03"1, P04'3 and TOC. Further, first flush effect was also estimated graphically. The cumulative volume/total volume vs. cumulative mass/total mass curve showed that the curves of many constituents ascended above the bisector line (45°), indicating aclear first flush effect. The chemical composition ofsedimentwas observed to display the order: TN >IC >TP >K>Ca >CI >Mg >Na. Ionic composition of the top soil of watersheds as well as the nature of the composite bonding of the sediments was considered to be the governing factors. High concentration of inorganic carbon (as displayed) could also possibly be related to the carbonate deposits in geological period of the area. The sediments from 2% slope watersheds displayed higher concentration of Ca, Mg, and CI as compared to other watersheds. Ionic composition of sediments was observed to be influenced by the nature of treatment. The concentration of trace elements for all watersheds was observed to display the order Mn > Fe > Zn > Cu > Pb > Cr > Cd. The trace elements were observed to display higher concentration than inorganic major ions, which apparently could be related to the ability of these metals to bond with fine sediment particulate and organic matter. The organic carbon bonding with sediments was observed to be greater in 2% slope watersheds. Generally, the sediments generated from watersheds having low runoff quantity, exhibited higher concentration of organic carbon than others. This could be due to the occurrence of organic carbon in the watershed soils and the available nutrients which were more favorable for soil biochemical activity and growth vi of flora and fauna.. The correlation among the ionic constituents of sediment exhibited good relation to each other except K+1 and inorganic carbon. The mass loads of sediment bound chemical constituents showed the order: O.C >TN >IC >Ca+2 >TP >K+1 >CI'1 >Mg+2 >Na+1 The control watersheds displayed highest loads in comparison to other watersheds. The computed values of site mean concentration (SMC) of the sediment generated from the watersheds increased with degree of slope of watersheds, clearly indicating importance of slope factor. Also, the SMC values decreased in watershed of both slopes with the scale of runoff treatment. The study year 2003 exhibited higher SMC as well as export coefficient values than 2002, apparently because of higher rainfall depth in 2003. Higher export coefficient (of sediment) values were exhibited for control watershed of both study slopes. Lower values of the computed loading rates were displayed by the 2% slope watersheds in comparison to 8% slope watersheds having similar runoff treatment. For the watersheds having same slope, the loading rates differed according to the scale of runoff treatment, with the higher scale showing the lower loading rates. The control watersheds exhibited highest loading rates in watershed of both slope categories but the percentage of increase in 8% slope watersheds was observed to be higher than in 2% slope watersheds, reflecting the impact of the runoff quantity. Non linear regression between SMC value and mean runoff depth of both 2% and 8% slope yielded asignificant power relationship for both slope (R2 =0.95, 0.92), respectively. Also, significant relation between load and mean runoff depth for both slope was obtained (R2 = 0.73, 0.83), respectively. Multiple regression analysis was attempted to estimate the relationship between loads and the causative factors. Following expressions emerged statistically significant: Sediment load (ton/ha) = 1.24log[Runoff (m3 )] -0.20 (R2 =0.78) Sediment load (ton/ha)= 122logf Runoff(m3 )] +1.09 log[A(ha)J -3.8 R2=0.83 Sediment load (ton/ha)= 1.21 logfRunoff(m3)J +1.07log[A(ha)]- 0.0043 (T)-3.77 R2 =0.92 Sediment load (ton/ha)= 1.23logfRun0ff(m3 )] +U2log[A(ha)]. 0.004 (T) +0.011(EI30). 4.14 R2 =0.98