Increasing the effectiveness of research on agricultural resource management in the semi-arid tropics of southern India by combining cropping systems modeling with farming systems research: A rewarding experience for Tamil Nadu farmers.

Introduction

Tamil Nadu Agricultural University (TNAU) with a research mandate to improve dryland cropping systems in Tamil Nadu, India, collaborated with ICRISAT and Agricultural Production Systems Research Unit (APSRU), Australia to combine cropping systems modeling tools with farming systems research expertise to deal with farmers' decision making process in managing dryland systems.

   
Map 1. The area covered in this study (green pattern in Coimbatore district map), situated in the southern state of Tamil Nadu (within pink border of India map) in India.

The major components of such cropping systems include cereals (sorghum, pearl millet and maize), legumes (groundnut), and commercial fiber crops (cotton). Sorghum and pearl millet growing is diminishing due to substitution by higher value alternatives. The remaining areas are used mainly for animal fodder and feed, and grown without any supplemental irrigation. Higher value crops are grown with supplemental irrigation but are increasingly subject to pest and disease problems. Farmers now need to deal with increasing complexity in their management and decision-making as a result of new cropping systems options, and there is a need for new tools that may help both the farmers and their advisors in the decision-making process. In this project, cropping systems modeling is being tested as a new tool.

 

Objectives

• To evaluate the capability of the cropping systems simulator, APSIM, to predict the performance of cropping systems in Tamil Nadu that includes cotton, groundnut, soybean, sorghum and maize.
• To test the idea that APSIM be used to enhance interactions between farmers, researchers and extension personnel to raise productivity and reduce risk in Tamil Nadu farming operations.

Approach to technology evaluation

The project builds on activities already under way in TNAU at its main campus in Coimbatore, and Aliyarnagar Agricultural Research Station (AARS). There are three components to the approach:

• Collation of results from past and ongoing field research throughout the Coimbatore/Tamil Nadu regions, focusing on studies where there is a balanced set of crop management, soil, weather and crop performance data have been collected that provide a suitable minimum data set for simulations.
• Evaluation, and where necessary, modification of APSIM module parameters (Ozcot-Cotton, Peanut, Soybean, Sorghum and Maize) to the point where users of APSIM have confidence in its application for this region.
• Development, adaptation and application of a methodology for integrating simulation in farmer decision-making, acceptance of the methodology by extension personnel, and teaching of the methodology in extension classes.

Data sets for APSIM Validation

Initially, TNAU, with help from ICRISAT and APSRU, upgraded necessary facilities – an automatic weather station was installed at AARS and in a farmers' field. Improved soil sampling and light interception monitoring equipment was acquired for sampling from farmer's fields. APSIM software configured for Tamil Nadu conditions was provided. Staffs at AARS have developed the capacity to use the APSIM-APSFRONT model interface to simulate the performance the desired crops, particularly groundnut. A long-term weather data set from AARS was compiled (1979-1999). Soils from farmers' fields were collected and analyzed. Minimum data sets for groundnut, cotton and sorghum of climate, crop growth and development, crop management and soil water were collected.

During the 1998-99 Rabi (post-rainy) season, three farmers fields in villages close to AARS were selected and monitored. An automatic weather station was installed in one field. The crop was sown in late December 1998 and monitored for phenology, crop canopy development and growth, and soil moisture through its growth. Data necessary to initialize APSIM were then used together with the weather data to compare the simulation with observed crop performance. APSIM accurately predicted the dates of flowering, podding and maturity, and also gave reasonable predictions of pod yield and vegetative biomass at maturity (Table 1).

Table 1. Experimental conditions observed and APSIM-model predicted phenology, pod yield and biomass for peanut cultivars (TMV2) in three villages (Reddiyarur, Angalakurichi and Avalchinnampalayam) in Aliyarnagar region of Coimbatore, Tamil Nadu, India.

Location	Sown (date)	Irrigation (mm)	Flowering (DAS)		Podding (DAS)		Maturity (DAS)		Pod yield  (kg ha -1 )		Biomass (kg ha -1 )
			Obs	Simul	Obs	Simul	Obs	Simul	Obs	Simul	Obs	Simul
Reddiyarur	20/12	325	37	35	66	64	90	94	2590	2670	4750	5200
Angalakurichi	21/12	400	37	37	66	66	94	94	2580	2700	4820	4900
Avalchinnam- palayam	20/12	350	41	39	68	65	89	92	3700	3700	4000	4000

APSIM modeled scenarios vs. Farmers' observed peanut yields

APSIM was used to simulate crop growth and yields for a range of possible sowing dates between October and February. Predictions were suggestive of changing the sowing date to mid-December, for possible increase in peanut pod yields considerably – an increase of about 1000 kg ha -1 would be expected (Table 2).

Table 2. APSIM-Peanut model predicted pod yield (kg ha -1 ) for different sowing dates.

Sowing dates	Peanut pod yield (kg ha -1 )
05 October	1390
20 October	1890
05 November	2090
20 November	2470
05 December	2560
20 December	3050
10 January	2300
25 January	2160
05 February	2000
Mean	2210
SD	431

Table 3. APSIM-Peanut model simulated pod yield (kg ha -1 ), as against observed yield from three different farmers' fields in three villages with varied sowing dates.

Farmer/village	Sowing date	Pod yield (kg ha -1 )		Biomass (kg ha -1 )
		Obs	Simul	Obs	Simul
Gopinath/Angalakurchi	05 December	1800	1900	6060	6018
RMSD		100		44
Angalakurichi	20 December	3450	3350	8390	9000
RMSD		94		609
Avalchinnam palayam	Late sowing	2190	2200	7090	6000
RMSD		10		1090

These predictions are supported by observed and simulated yields for groundnut from three farms – these simulations indicate that mid-December sowing of peanut gives higher pod yields (Table 3).

   
Figure 2. Scientists' visiting experimental plots in farmers' fields, and comparing the differences in peanut growth as affected by varied sowing dates at Avalchinnam palayam village in Tamilnadu.

Conclusions

This research is of immediate value to the farmers in the Pollachi region of Tamil Nadu, where farmers restrain themselves from adopting new crops and technologies, with which they have little experience. Computer simulation of crops can now be used to help farmers to identify strategies to make best use of scarce water resources, and in this case to avoid low temperature stresses. The computer modeling can substitute for a significant part of experimentation. By using the model while interacting with the farmers, the researchers and farmers were able to gauge the reliability of the model. When the model accurately simulated the yield of the farmer's crop, farmers became interested in using the model to explore other scenarios. Our experience working with groundnut-growing farmers in the Pollachi region was expanded by comparing sowing dates, cultivar choice and fertilizer rates. The first important step was to make observations on the farmers' fields – monitoring soil water, nitrogen, crop growth and yield. This comprehensive collection of high quality data provided information for validating the model, and also attracted the interest of farmers in making field observations that they had not previously done. The most successful exercise was the examination of the question of whether to sow the groundnut crop in December or January. The model predicted that December sowing resulted in greater yield than January sowing. The model predictions were then confirmed by the farmers who now expect an increase in pod yield by 1000 kg ha -1 than with early sown peanut crop. Use of computer simulations proved to be a rewarding experience for farmers, researchers and extensionists. Now, the approach is being extended to larger areas in pollachi region of Tamil Nadu state in India.

Collaborator

Mr. R. Madhiyazhagan, Associate Professor (Agronomy), Aliyarnagar Agricultural Research Station (AARS), Aliyarnagar, Tamil Nadu Agricultural University, Coimbatore, India.

For more information please contact:

Mr. R. Madhiyazhagan
Associate Professor (Agronomy),
Aliyarnagar Agricultural Research Station (AARS),
Aliyarnagar, Tamil Nadu Agricultural University,
Coimbatore, India.

and

Dr. R. J. K. Myers
Principal Soil Chemist,
ICRISAT-Bulawayo, Zimbabwe.