1. To intensify research on aerobic meteorology, hydrometeorology, micrometeorology, crop production meteorology (crop-weather studies), soil climatology, microbial meteorology, herbicide meteorology, forest meteorology, horticultural meteorology, animal meteorology, seed meteorology and other related subjects.

  2. To provide center of activities for the different agromet observatories of TNAU.

  3. To strengthen the agricultural meteorology education for the State of Tamil Nadu through

    Certificate course for B.Sc (Ag) graduates on agricultural meteorology.

    Specialized degree course for agricultural meteorology.

    Short course for farmers of Tamil Nadu.

  4. To improve weather forecast in collaboration with NCMRWF, New Delhi.

  5. To develop resource personnel on drought climatology, agromet-spectral modeling for yield forecast, Weather modifications and allied techniques.

  6. To develop simulation crop modeling and as well as crop pests - weather interaction modeling.

  7. To undertake climate change studies.



Rice grain yield of both kuruvai and Thaladi seasons over 30 years (1961-1990) were correlated with respective weather data to find out the crop weather relationship of concerned season. The study indicated that reproductive stage of rice was very critical for both kuruvai and Thaladi seasons for weather parameter. In addition maturity stage of kuruvai season and vegetative stage of Thaladi season were also critical to weather. During Thaladi season, correlation study indicated that positive correlation was observed for maximum temperature at vegetative and reproductive stages. 

Maize (Rainfed):

The experimental result revealed that pre monsoon sowing with recommended plant population (60 x 20 cm) recorded higher grain yield (2970 kg/ha) compared to other treatments studied. Among N fertilizer application, 50 per cent basal +50 per cent flexible as per the rainfall receipt recorded maximum yield (2510 kg/ha) and it was 42 per cent higher than the full basal application.

Cotton (Irrigated):

In the crop weather relationship study in cotton, sowing cotton in the first week of August second fortnight recorded maximum yield of 1357 kg/ha. When the sowing was delayed up to September first fortnight, the yield reduction was 89 per cent. If the sowing was further delayed up to September second fortnight, the yield reduction was 143 percent. If the sowing was done in earlier July, the yield was declined, but it was not significantly different from August second fortnight sowing.


Coimbatore lies on the rain shadow region of western ghat in India.  Coimbatore gets most of its rainfall from northeast monsoons (October – December). The strong Asian summer monsoon is generally associated with positive sea surface anomaly in the western Pacific (El-Nino).  In the present study an attempt was made to categorise the rainfall data of Coimbatore based on El-Nino episodes. The results revealed that El-Nino favoured both annual as well as northeast monsoon rainfall while the La-Nina had influenced the South west monsoon rainfall.  The study also indicated that in Coimbatore during El – Nino years crop production is with less risky during northeast monsoon season and the farmers can choose even crops with comparatively high water requirement, whereas in La-Nina years higher yields could be reaped from the southwest monsoon season in certain taluks of Coimbatore district, India.  


By using historical rainfall data an integrated study was attempted between the Length of Growing Period (LGP) and Global weather phenomenon factors like El-Nino and La-Nina.  Successful crop production and choice of crops based on the length of growing period is highly influenced by the El-Nino and La-Nina conditions.  The method suggested by Jeevananda Reddy (1983) was used to find out the LGP.  El-Nino, La-Nina and Normal years were disintegrated from the total years of study by collecting the information from the web site of Bureau of Meteorology.  Eventhough not much difference could be observed from the length of growing period between various conditions studied, the occurrence of intermittent wet spell and dry spell varied and that decides the yield of the crop.  During El-Nino years crop production is less risky with the occurrence of continuous wet spell as compared to the La-Nina years.  Based on the forecast of the El-Nino or La-Nina situations, agronomical practices such as land configuration, selection of crops and varieties and other management practices can be manipulated to suit the expected weather challenge. 


Indigenous knowledge and beliefs of the farmers many times become true as far as weather forecasting is concerned.  Almanac is one such source that provides valuable forecast information and many people believe on that.  Melnokku nal or keelnokku nal was mainly based on the position of sun from equator either to tropic of cancer or to tropic of Capricorn and also due to position of moon around the earth. A study was undertaken to find out the scientific rationale behind the melnokku nal and keelnokku nal information given in the Almanac.  The study indicated that in respect of rainfall in all the keelnokku nal of different months there was rainfall irrespective of the quantity received while in melnokku nal of different months rainfall was absent in February and March months. Relatively higher rainfall was recorded in keelnokku nal during April, May, June and October months and reverse trend for melnokku nal in the remaining months of the calendar year.


Spodoptera is a polyphagous pest and its activity is highly influenced by the prevailing weather conditions. A study was conducted at TNAU for two years (1999/ 2000) to find out the influence of weather conditions that lead to sudden eruption of this pest by collecting the moths trapped in pheramone trap. Frequency analysis of Spodoptera moth population revealed that maximum temperature of 26°C, minimum temperature of 18°C favoured moth activity.  Increasing relative humidity and cloudiness increased the moth population.  Emergence of more than 50 moths per week could be observed when the morning relative humidity increased from 85.6 to 89.4 per cent.  At rainfall of 20 to 40 mm/week the moth population was higher.  Different seasons also had influenced the moth population and South West Monsoon season had more influence in erupting the population than other season.  The derived information could be well utilised for forecasting the pest incidence, monitoring its occurrence  and control of this polyphagous pest.


Based on the interaction between earth and moon in relation to Sun, there are two phases of the moon viz., full moon and new moon.  The shadow effect of earth triggers new moon (Ammavasai) while the opposite one is full moon.  Each month is governed by both new moon and full moon and in between these two there are fourteen thithies covering the 14 days interval and those are Prathamai (1), Thuthiai (2), Thiruthiai (3), Charthurthi (4), Panchami (5), Sasti (6), Sabthami (7), Astami (8), Navami (9), Dhasami (10), Yegadhasi (11), Thuvadhasi (12), Thiriodhasi (13) and Sathurthasi (14).  A study was undertaken to find out the relationship between the rainfall and the different thithies occur in between the two phases of the moon viz., full moon and new moon  and the results revealed that the first eight thithies succeeding new moon (ammavasai), and eight thithies preceding to new moon did influence annual rainfall events.  Further it is observed that with reference to higher intensity rainfall, it occurred normally the eight thithies preceding to new moon as compared to thithies succeeding to new moon.  Almost similar results could be noticed for both South west monsoon season and for North east monsoon season. Over year analysis indicated that towards full moon phase (Valar Pirai), the thithi Sasti had a tendency to offer high rainfall while such effect was offered by Yegadasi thithi towards new moon. Neverthless, all the thithies both during new moon phase and full moon phase did have rainfall events though not as high. The results further indicated that high intensity events occurred frequently during new moon phase as compared to full moon phase.



The Southern Indian Peninsular receives monsoon rain in summer (June-September) and winter (October-December) that is highly variable. The ability to predict seasonal rainfall and associated crop yield distributions in advance could have a major impact on the management of dryland cropping systems in this region. A comprehensive systems approach and a climate forecasting system were used to quantify the cropping system management responses to seasonal climate forecasts. In this part of India, the SOI is positively related to summer monsoon rainfall, but exhibits an inverse relationship for the winter monsoon. The probability of exceeding the median summer monsoon rainfall following a warm ENSO phase in April/May is 31%, while it was 60% in cold phase years. Conversely, warm phase during June/July showed 82% chance of exceeding median rainfall during winter monsoon, while a cold phase showed only a 28% chance.  This finding is consistent with general scientific understanding of the ENSO (El Niņo Southern Oscillation) life-cycle.

The yield and gross margin of peanut (summer monsoon crop), sorghum (winter monsoon crop) and cotton (overlapping in to two seasons) based cropping systems, showed significant response to seasonal climate forecasts. The peanut yield potential cold phase is increased by 374 kg/ha (58%) compared to warm phase season type. The higher yield potential during the cold phase years are caused by more opportunity of planting rains and greater in-season rainfall. The yield of winter monsoon sorghum after summer peanut can be substantially increased during warm ENSO phase years by 36% and 68%, respectively over climatologically average yield. Our analysis indicated that forecast responsive sorghum crop management strategies were superior to traditional management (p<0.01). The chance of producing sorghum fodder for animal is lower in cold phase years than warm phases. The farm management decision on choice of cropping system (peanut-sorghum and cotton) before start of the season also showed significant response to seasonal climate forecasts. Cropping system simulation showed that during cold phase during April/May, peanut-sorghum systems were superior to cotton systems, while following warm phases, cotton often benefited from greater winter monsoon rainfall coinciding with critical crop growth period. The results demonstrated a significant response of dryland cropping system management decisions to climate-forecasting system. The simulated crop yield and gross margin probability distributions conditioned by climate forecasts can support the small-holder farmers to manage the climate risk in cropping systems through enhanced decision-making capacity of farmers.


Agricultural systems in semi-arid tropics are strongly affected by climate variability. Timing and frequency of rainfall events conditioned by the inter-annual large-scale ocean-atmosphere circulation features strongly influences the dryland cropping systems. Very low level of rainfall use efficiency has been observed in these systems due to inappropriate crop management decisions. However, the current emerging capacity to forecast the likelihood of future climatic distributions can enhance the rainfall use efficiency of smallholder cropping systems exposed to climatic risk.  

Global climate models (GCMs) have done a reasonable job in simulating many aspects of the climate variability on inter-annual time scales. Climate simulations through global circulation model, downscaled for selected locations of the southern peninsular India-showed reasonable prediction. Briefly, a statistical transformation of seasonal rainfall output fields from each of several GCMs were used to identify optimal predictors. Model Output Statistics (MOS) and strong multivariate techniques have been used to generate dominant large-scale modes of patterns in the simulated GCM fields. The monthly rainfall hindcasts of GCMs were disaggregated into daily values (five realizations) using a stochastic weather generator. Both predictor selection and regression were fully cross-validated to avoid false skill in rainfall prediction. The observed and hindcast rainfall showed significant relationship (r=0.512).

The simulated rainfall was combined with the systems simulation approaches to find out the optimal management practices for improving rainfall use efficiency of dryland cropping systems. The groundnut yields simulated with observed weather and stochastically-disaggregated monthly hindcasts also showed reasonable closeness (r = 0.554). Emphasis was given to assess the impact of El-Nino Southern Oscillation (ENSO) related inter-annual climate variability on ground water resources. Additional analyses examined crop evapotranspiration and irrigation requirement for maize crops in the study region.  Results indicate that effective rainfall was 18% less in warm ENSO years that also exhibited 14% greater crop evapotranspiration, necessitating supplemental irrigation. Such information on the type of water requirement conditioned by ENSO phases would assist farmers in making water management decisions.

Development of weather based Pest and disease forecasting in cotton

Weather pest relationship studies taken up in cotton crop showed that there is a positive relationship existed between aphid incidence and relative humidity (morning, evening and mean), temperature (maximum and minimum) and rainfall. In contrast, a negative relationship existed with solar radiation bright sunshine hours and wind speed in both early sown and late sown cotton crops. Adult moth catches in pheromone trap indicated that there is positive relationship between maximum temperature, minimum temperature, afternoon relative humidity and Helicoverba moth emergence. Negative relationship could be observed for rainfall, solar radiation, bright sunshine hours and wind speed. 

Development of weather based forewarning system for groundnut pest and disease 

The observation on rust and late leaf spot was studies at TNAU, Coimbatore at three days interval in all the three dates of sowing viz., normal date of sowing, 15 days after normal sowing and 30 days after normal sowing. In all the three sowings there was pest control, disease control, pest and disease control and total unprotected control. In the first sowing the rust incidence was upto 65.33 per cent as against 37.33 per cent in disease control plot. In second sowing the rust incidence was upto 71.78 per cent as against 42.67 per cent in disease control plot. In third sowing the rust was 64.44 per cent against 43.78 per cent as against 46 per cent in disease control. In second sowing the incidence was upto 68.44 per cent as against 42.67 per cent in disease control. In the third sowing the disease incidence was 60.44 per cent as against 42.0 per cent in disease control plot. The result indicated that the rust incidence was severe in second date of sowing (4.12.2001) as compared to first sowing (22.06.2001) and third sowing (25.07.2001). Such trend was not observed for late leaf sport disease. 

The data revealed that irrespective of two sowings (I sowing 22.6.2201, II sowing 4.7.2001) the pod yield was similar for control, but however it varied for other treatments between two sowings. In respect of first sowing highest pod yield was obtained under pest and disease control as compared to pest control, disease control separately, such difference was not noticed in respect of second sowing. The relation between weather elements and pest and disease incidence was established through correlation and regression. 

Forecasting NEM rainfall (SCF) and assessing its validity with real occurrence

From the rainfall records of different stations of Tamil Nadu. The El-Nino years rainfall data alone were separated and analysed.

South west monsoon:

Analyses of long term average of South west monsoon rainfall during El-Nino years revealed that during El-Nino years, in general the amount of rainfall decreases when com-pared to the South west monsoon normal rainfall of all years. Only North eastern parts of Tamil Nadu received more than 400 mm of rainfall in SWM season during El-Nino years.

North east monsoon:

Analyses of long-term average of North east monsoon rainfall during El-Nino years revealed that in many places, the quantum of rainfall increased during the El-Nino all the places in Tamil Nadu received more than 300 mm of rainfall during NEM. Coimbatore, Erode and Salem districts received 300 – 400 mm of rainfall. Southern districts received 400-500 mm rainfall and the places along the coastal line received more than 600 mm rainfall. By employing Australian Rainman software seasonal climate forecast was developed for Southwest monsoon and Northeast monsoon 2001 for different centers and communicated. Informations are given with actual data. 

Validity of medium range weather forecasting under farm situations 

The highest accuracy in rainfall prediction was observed in the month of March, 2001 (96.3%) followed by February, 2001 (95.7) and January, 2001 (84.2%). Lower RMSE values indicate better predictability. Lower RMSE values were observed for January, February and March, 2001 indicating better prediction. Higher ratio score indicate good prediction. Ratio score was >0.8 for the months December, 2000 and from January to March 2001. Seasonal analysis of rainfall data indicate that the prediction was better in Cold Weather Period, 2001 followed by Hot Weather Period, 2001. North East Monsoon, 2000 and South West Monsoon, 2000.

Maximum Temperature:

Monthwise analysis of maximum temperature indicate that in all the months the prediction was correct and usable for > 70% of the occasions except June and July, 2000. All the months showed lesser RMSE values indicating the effectiveness of the predicted data. Correlation between the predicted and observed data indicated that the correlation values were more than 0.4 for all the months except June, July, December, 2000 and for March, May and June, 2001.

Minimum Temperature:

Analysis of minimum temperature indicated that in all the month the prediction was correct and usable for > 80% of the occasions except September, 2000 January, 2001. All the months showed lesser RMSE values indicating the effectiveness of the predicted data. Correlation between the predicted and observed data indicated that the correlation values were comparatively lower in June-September, 2000 and in January and May, 2001.

Wind Speed:

Maximum wind speed was predicted and observed during the month of July 1999, predicted wind speed was cent per cent correct and usable during February and March, 2000. Except in May and June, 2000, prediction of wind speed was comparatively satisfactory. Seasonal analysis of wind speed indicated the forecast accuracy was higher during CWP, followed by NEM, HWP and then SWM. In general prediction of wind speed was correct and usable for > 80% in all the seasons except SWM. 

Wind Direction:

Month-wise analysis of wind direction indicated that the prediction of wind direction was not good during the months between February and May, 2000 as well as during October, 1999. Correlation analysis indicated that during November and December, 1999, the correlation value between the predicted and observed wind speed was >0.9. Seasonal analysis of wind direction indicated that during NEM and SWM, the prediction was satisfactory as compared to the other two seasons.

ENSO – A viable predictor of Tamil Nadu Seasonal rainfall

The impact of climate variability on agriculture is large, particularly in those countries impacted by the El Nino / Southern Oscillation (ENSO) phenomenon such as Australia, India, Indonesia and Africa.  An attempt was made to document the relationship between the ENSO and rainfall of different locations of Tamil Nadu and using this relationship monsoon rainfall was forecasted and verified. The results are summarized below:

  • The spatial coherence of the correlations with the Southern Oscillation is encouraging. During SWM, there is a consistent positive correlation existed in most of the study locations, in the sense that monsoon rainfall is higher (lower) during a La Niņa (El Niņo), while there is a consistent negative relationship existed with NEM rainfall.  The sign actually changes from positive to negative between September and October, so that the individual monthly rainfall for October is correlated in less magnitude with the SOI.  This presents problems for crop management in Tamil Nadu in the month of October, which is the critical time for sowing of North east monsoon crop.

  • NEM Rainfall of most of the places in Tamil Nadu deviated in positive side during El-Nino years and on negative side during La-Nina years.

  • South west monsoon onset was delayed by 34.7 per cent during El-Nino years, while it was only 13.3 per cent during La-Nina years.  During 80 per cent of the La-Nina years, the monsoon onset was on normal date.

  • In 87.1 per cent of the El-Nino years and in 90.2 per cent of the normal years, the North east monsoon onset was either early or normal.  During El-Nino year, farmers of Tamil Nadu can take up pre-monsoon sowing during NEM period to utilize the full benefit of the rainfall.

  • Both El-Nino and La-Nina had good influence over withdrawal of NE monsoon favourably.  More than 95 % of the El-Nino and La-Nina years had either normal or late withdrawal which was favourable for agriculture and the crops with escape from the terminal season drought. 

  • Under El-Nino situation, the mean length of growing period was 21 weeks.  Hence, if the expected situation is an El-Nino year, farmers can go in for more water requiring long duration crops. During La-Nina years, although the length of growing period was more, the crop would suffer due to intermittent dry spell, as there were only four wet spell weeks and nine dry spell weeks with in the growing period. In normal years, the growing period registered was 16 weeks and a medium duration crop may be grown with out any risk under normal years by providing necessary drainage facilities.

  • The seasonal rainfall forecast accuracy ranged from 47.8 to 85.7 % over different seasons. 



A study on assessment of climate change impacts on cropping systems in southern India has been undertaken. Climate change scenarios were developed by following two methods (i) fixed climate change scenario and (ii) daily outputs of a Regional Climate Models (HadRM2). Validated crop models were parameterized for current management practice.

The simulated climate by HadRM2 under control (CTL) was compared with observed climate (base line) for past three decades to understand the likely pattern of current climate in future (2041-2060) with 1990 GHG emission scenario. The temperature predictions are very close to current observations. The simulated annual rainfall under CTL was consistently lower by 16.8% compared to current observed rainfall. The difference in rainfall might be unimportant for irrigated rice crop performance, but will have major effect on rainfed crops. The future climate change scenario (GHG) of HadRM2 showed increase in solar radiation, maximum temperature, and minimum temperature by 0.35 MJ/m2, 3.4° C and 3.6° C, respectively and decline in rainfall by 9.5% over control (CTL), corresponding to the period 2041-2060. The temperature scenario projected by the model is higher by 1.5° C greater than the pessimistic scenario of earlier reports for south Asia. This uncertainty in temperature could produce a larger difference in impact assessment and needs careful interpretation.

Rice based systems of southern India exhibited greater probability of yield reduction under future climate change scenario with temperature and CO2 increments and HadRM2 model projections under GHG scenario. The average decline in rice yield is –8.4% and -7.6% for the first (Jun-Sep) and second season rice (Oct-Jan/Feb) of double crop systems, while the average reduction for single crop rice (Sep-Dec/Jan) system is lowest with -3.7%. Impact of climate change simulated by HadRM2 has revealed that the rice yield reduction under GHG scenario ranges from 370 kg to 783 kg/ha with current level of management. Delay in planting of first season rice by a month (July planting) reduced the rate of yield decline. However, delayed planting of first rice by a month will have greater overlap with the subsequent seasons, making the options very difficult. The economic inequality gap widens among the farmers due to varied level of management under future climate change. Farmers have already exploited the production potential by shifting the time of planting, leaving very little scope for change in practice. Researching for high temperature tolerant and drought resistant rice verities might be the better adaptation option.



For further information


For Comments and Suggestions