Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Inde → 2005 → Modelling of Drought Under Different Agro-Climatic Zones of Gujarat

Maharana Pratap University of Agriculture and Technology (2005)

Modelling of Drought Under Different Agro-Climatic Zones of Gujarat

Deora, Bhom Singh

Titre : Modelling of Drought Under Different Agro-Climatic Zones of Gujarat

Auteur : Deora, Bhom Singh

Université de soutenance  : Maharana Pratap University of Agriculture and Technology

Grade : Doctor of Philosophy (PhD) 2005

Résumé présenté dans KrishiPrabha (Indian Agricultural Doctoral Dissertations Repository)

Modelling of drought is one of the most important aspects in the rainfed farming. The present study was undertaken to model drought in different agro-climatic zones of Gujarat, by analysing rainfall and estimating dry spells and water deficits, determining suitable frequency distributions for extreme weekly water deficit, developing and validating appropriate artificial neural networks and stochastic models, studying drought spread using geostatistical modelling and suggesting suitable crop planning. Weekly meteorological data for 22 to 30 years (upto 2003) and soil data for sixteen stations (representing eight agro-climatic zones), namely Kothara, Radhanpur, Rajkot, Amreli, SK Nagar, Khedbrahma, Arnej, Dhandhuka, Junagadh, Mahuva, Anand, Nawagam, Bharuch, Surat, Navsari and Gandevi were collected. Meteorological and soil data for Dahod were also included for geostatistical analysis.

Weekly rainfall data were analysed for onset and withdrawal of rainy season and determination of dry spells. Results of rainfall analysis of different agro-climatic zones indicated that average annual rainfall varied from 368.74 mm at Radhanpur to 1580.53 mm at Gandevi with coefficient of variation of 54.94 and 32.25 per cent, respectively. Mean seasonal rainfall received from June to September ranged between 337.68 mm at Kothara to 1541.33 mm at Gandevi. Maximum and minimum mean weekly rainfall during monsoon season ranged from 46.91 to 197.55 mm and 2.10 mm to 42.48 mm, respectively in different agro-climatic zones. Annual, seasonal and weekly rainfall analysis of different agro-climatic zones indicated that the coefficient of variation at Kothara is highest, signifying extremely erratic rainfall distribution compared to other stations. Effective monsoon onsets from 25th to 28th week and terminates from 36th to 40th week in different agro-climatic zones of the state. Dry spell analysis indicated that mid season droughts are likely to occur with 26 to 50 per cent probability at Junagadh, Anand, Nawagam, Bharuch, Surat and Navsari ; with 51 to 75 per cent probability at Rajkot, Amreli, SK Nagar, Khedbrahma, Arnej, Dhandhuka and Mahuva, and < 25 per cent probability at Gandevi. Kothara and Radhanpur experienced mid season droughts on regular basis with high probability. The possibilities of late season droughts during maturity period are high (> 50% probability) at all the stations except Gandevi (< 25 % probability).

Under normal condition, moisture status at all the stations situated in half of the eastern part of the state is sufficient for satisfactory growth of rainfed crops, whereas, all the stations located in the other half (western part) experienced moisture deficit during the crop growth period. Drought evaluation study indicated a frequency of one drought year in every two to three years. The analysis also indicated that drought does not occur simultaneously in all the agro-climatic zones but usually occur sporadically all over the state. However, during some years, all the zones experienced droughts of varying intensity. The results indicated that droughts are common not only in arid and semi-arid climate but equally occur even with higher frequency in dry sub-humid climate also.

Weekly reference evapotranspiration and water balance components were also determined. Frequency analysis of weekly maximum water deficit during crop growing season indicated that Gumbel distribution fitted best at Kothara ; Weibull (maxima) distribution at Dhandhuka, Junagadh, Mahuva, Nawagam and Surat and log Pearson type-III is the best probability distribution at rest of the stations.

The artificial neural network (ANN) generated water deficit and observed water deficit have indicated similar trend for all the weeks for different stations. Highly significant values of correlation coefficient of the regression analysis for different stations indicated that the network has accurately estimated the water deficit and the provided 7-4-1 architecture is adequate for the network. The trained network model during validation has predicted water deficit reasonably accurate for all the stations.

Developed stochastic models were found to be adequate for forecasting the weekly water deficit for two years ahead (2002 and 2003) as seen from the highly significant values of correlation coefficient between observed and predicted series. Spatial and temporal variability of water deficit during 1987 drought year using geostatistical modelling technique indicated that the occurrence and spread of drought has a tendency to originate in the water deficit region, spreads in the direction of similar deficient regions and then dissipates.

Mono-cropping as well as double cropping systems are possible in all the agro-climatic zones of Gujarat (except hot arid and part of semi-arid areas), mainly due to presence of heavy soils. Short duration rainfed kharif crops can be successfully grown in different zones. Long duration kharif crops experience water stress for short period during their critical growth stages. However, sustainable production can be maintained by providing protective irrigation through harvesting of surplus water during rainy season.

KrishiPrabha (Indian Agricultural Doctoral Dissertations Repository)

Page publiée le 27 janvier 2015, mise à jour le 30 janvier 2021