INTRODUCTION Farmers in tank fed irrigation used to encounter different income related risks in their occupations. To compensate the income loss, government encourage these farmers to enroll in insurance scheme like Pradhan Mantri Fasal Bima Yojana (PMFBY). Crop insurance is one of the options for farmers to manage the monitory loss due to yield loss. It contributed in stabilisation of farm production and income due to yield loss faced by the farming community. An insurance programme is effective when it induces farmers to take on more risks that they would not have taken otherwise. Farmers will take on more financial risks as insurance reduces risk in farming, a practise known as 'risk balancing' (Liang, 2014). But, recent studies indicated that the compensation measures of crop insurance coverage are very low and less than 5.00 per cent in India. For paddy and groundnut crops, the proportion of farmers reporting crop loss is significant (more than 25%), but insurance coverage is inadequate (Aditya et al., 2018). Farmers earlier who were enthusiastically participating in the scheme through paying premium now are reluctant to pay premium for two reasons: first, there was delay in releasing the compensation from the insurance companies. Second the amount of compensation measure paid was inadequate. Hence, an attempt has been made to measure the total income loss of farmers when they encountered different risks. To quantify the risks Analytical Hierarchy (AHP) method was employed. RESEARCH METHODOLOGY The analytic hierarchy process (AHP) also known as multi-criteria decision-analysis method introduced by Satty (1980) was employed for construction of estimation of income risk. It is widely applied in outstanding works of various fields relating to best option selection, conflict solution, resource allocation and optimization of the decision-making process. In this study, the AHP is employed to establish actual contribution of the components and sub components in estimation of income risk through the steps given below and followed by (Rajeshwaran et al., 2021 and Ejovi et al.. 2021). The basic procedure follows for the AHP: 1. Hierarchy construction 2. Developing a pairwise comparison matrix for each criterion 3. Normalizing the resulting matrix 4. Averaging the values in each row to get the corresponding rating 5. Calculation and checking the consistency ratio 1. Hierarchy construction Hierarchy is established by breaking down the overall goal that is measures of estimation of income risk through basic elements. These fundamental elements of hierarchy construction are divided into two categories based on occupation: farming + livestock and farming + wages. Further, the component, farming risk, is categorised into 5 different sub-components; they are production risk, marketing risk, financial risk, human resource risk, and institutional risk; and another component, livestock risk, is categorised into three sub-components; they are production risk, marketing & financial risk, and institutional risk (Roger et al., 2011). The review of literature and authors’ critical judgments has led to the construction of the hierarchical model consisting of different components. Making such construction helps to identify the components with their exclusive domain. Moreover, it helps to control the duplication of subcomponents. 2. Developing a pairwise comparison matrix for each criterion. The contribution of one component over another component has to be measure through psychological scale that build up through psychological continuum, by ordering the components through psychophysical method. In this way the weight score of pairwise comparison was assigned with a scale of 1-9 as depicted below. Two items are equally important one item is extremely favoured to another Here, the weightage score assigned by the different experts are pooled together and average score of pairwise was worked out. The pairwise score were depicted in the matrix format Matrix of pair wise element = [█(├ █(C_11 C_12 C_13@C_21 C_22 C_23@C_31 C_32 C_33@C_41 C_42 C_43@C_51 C_52 C_53 )] )┤ Sum the values in each column of the pair wise matrix=∑_(i=1)^n▒C_ij 3. Normalizing the resulting matrix. To get the overall importance of one element over another element was worked out for average score of pairwise items in normalized matrix. To generate a normalised pairwise matrix that each element in the matrix was divided by its column total. x_ij=C_ij/(∑_(i=1)^n▒C_ij ) 4. Averaging the values in each row to get the corresponding rating. The sum of the normalized column of matrix was divided by the number of criteria used to generate weighted matrix. Moreover, this average score gives the percentage contribution of particular element towards the goal. w_ij=(∑_(j=1)^n▒x_ij )/n [█(W_11@W_12@W_13 )] 5. Calculation and checking the consistency ratio. Judges may assign the pairwise matrix scale without due consideration of relative importance of each elements. If it is so, the score one got in the early steps may not reflect the reality. To get the validity and reliability of the score, consistency check has to be carried out. The consistency ratio was calculated to make sure that the original preference ratings were consistent. There are 3 steps to arrive at the consistency ratio: 1. Calculation of consistency measure. 2. Calculation of consistency index (CI). 3. Calculation of consistency ratio (CR). Calculation of the consistency measure. To calculate the consistency measure, the matrix multiplication function =MMULT() is used for actual rows with average column. 1. Consistency measure is calculated by multiplying the pairwise matrix by the weights vector [█(C_11 C_12 C_13@C_21 C_22 C_23@C_31 C_32 C_33@C_41 C_42 C_43@C_51 C_52 C_53 )]*[█(w_11@w_21@w_31 )]= [█(C_(M_11 )@C_(M_21 )@C_(M_31 ) )] 2. Consistency vector is calculated by dividing the consistency measure with average criterion weight. C_(v_11 )=1/w_11 [C_11 w_11+C_12 w_(21 ) C_13 w_31 ] C_(v_21 )=1/w_21 [C_21 w_11+C_22 w_(21 ) C_23 w_31 ] 〖 C〗_(v_31 )=1/w_31 [C_31 w_11+C_32 w_21 C_33 w_31 ] 3. λ was calculated by averaging the value of the consistency vector λ= (∑_(i=1)^n▒〖cν_ij 〗)/n Calculation of the consistency index (CI). It was calculated by using the formula given below. Cl= (λMax-N)/(N-1) λ Max = averaging the value of the consistency vector N = Number of criteria Calculation of the consistency ratio (CI/RI where RI is a random index). It was done by following the formula given below. CR= Cl/RI CI = Consistency index value RI= Table value Random Index (RI). The RI was obtained from the random inconsistency indices given by Satty (1980) which is furnished below. Study area. Small and marginal farmers in tank fed irrigated areas comprise the universe of this study. The multistage random sampling method was used to select the ultimate sampling units. The study was conducted in Sivagangai district of Tamil Nadu. Which is having the highest net area of 63,749 hacters covered under tank irrigation compare to other districts of Tamil Nadu. Further, most of the small and marginal farmers were holding either livestock or wages along with farming as occupation. So, it was decided to choose 40 respondents from farming + livestock and 40 respondents from farming + wages occupations are selected. Thus, total 80 respondents from the district are selected for this study. RESULT AND DISCUSSION To estimate the income risk faced by small and marginal farmers, economic losses incurred in yield loss was worked out. In this regard, data pertaining to potential income earned by the small and marginal farmers are derived from reliable records such as Directorate of Economics and Statistics, Estimation of Cost of Cultivation / production & Related data from 2013 to 2018 and TNAU Agri portal. The data were further verified in triangulation with extension functionaries such as state officials and progressive farmers during the course of survey. The actual income earned by the small and marginal farmers in an annum was worked out from the data collected from respondents. The difference between the potential and actual income considered as the loss incurred by the farmers due to occurrence of risk events. This is presented in the Table 1. The data in the Table 1 depicts the estimated loss of income incurred by the farmers in tank fed irrigation system. It could be understood that the major income loss happened in livestock (Rs.26845) followed by farming (Rs.13483) and wage employment (Rs.9700). After arriving the total loss, it was decided to estimate the actual contribution of individual risk events through Analytical Hierarchy Process (AHP). Though in early days AHP (Khwanruthai, 2012) was used for decision making by prioritizing the best course of action from the available alternative through paired comparison method, in recent years this tool is employed for measuring estimations (Rajeshwaran et al., 2021). This tool is particularly useful when there is mixing up of objective value as well as judgement values to prioritize the events. Here adhering the methodology, analytical structure was worked out separately for two different categories of small and marginal farmers based on composition of occupations they were involved. In the first step, the loss incurred in two different occupations in proposition was calculated. In the second step of hierarchy construction, different weightage scores were assigned to different category of risks such as production, marketing, financial, human resources and institutional risks. Further, in third step the major risks identified related to the different categories of risks were put forth based on judgement of judges. Finally, based on the weightage score assigned by the judges the actual contribution of loss incurred was estimated in terms of money value which is presented in the Table 2 and Fig. 1-2. From the Table 2, it could be ascertained that farmers who are holding livestock as occupation are losing Rs.26845 from their annual income. This estimated income loss are happening due to poor-remuneration through selling of livestock during pandemic period followed by diseases and unprecedent death of livestock through to the tune of Rs.9210.94, Rs.4422.74, and Rs.3197.35 respectively. Regarding farming, farmers were losing their income mostly related to encounter of production risks followed by marketing risks. The production risks like escalation of inputs cost, high hiring cost of farm machineries, incidence of pest and diseases have dominating with the estimated losses of Rs.2117.25, Rs.1563.33 and Rs.1351.67 respectively. Through the escalation of cost of labours, farmers loss an estimated amount of Rs.1402.97 followed by selling the produce in lower price (Rs.1188.03). From Fig. 1&2 indicate that small and marginal farmers were losing income of around Rs.10000 per annum in wage employment which receiving of less wage followed by delay in payment and unable to get steady employment have contributed to the estimated loss of Rs.4475.33, Rs.2672.56 and Rs.1433.28 respectively. Further, the compensation paid and the compensation yet to be paid against the loss also was worked out from the primary data collected from the farmers and the consolidated account is presented in Table 3. From the Table 3, it can be understood that an average amount of Rs. 4145.13 was paid through crop insurance for the year 2020 against original loss of income of Rs. 13483 from farming sector. Hence, it is recommended and additional amount Rs.9338 yet be paid as compensation by considering the risks peculiar to small and marginal farmers. Similarly, in rising of milch animals own amount of Rs.15748 was paid as compensation for risks encountered in the occupation against the actual loss of Rs. 26845. Hence an amount of Rs.11097 may be paid to small and marginal farmers for the peculiar risks faced by them. If the farmers get that money from compensation it leads to increase the income level of small and marginal farmers in tank irrigated system (Tlholoe, 2016).