Estimation of Acreage & Crop Production through ...

4 downloads 56 Views 772KB Size Report
District wise acreage estimation of Soybean was done by using the following mathematical .... Ujjain and Sagar are the two other districts with an average yield ...
Estimation of Acreage & Crop Production through Remote Sensing & GIS Technique Abhishek Kumar Maurya

Abstract: Soybean is an important crop in Madhya Pradesh. The crop is grown in average 21.88 lakh hectare areas. The production and area can be estimated in a very concise method and in very less time with maximum accuracy. Remote Sensing technique provides a methodology to map areas of Soybean field with the help of MODIS (TERRA) Satellite data and GIS (Geographical Information System) database. MODIS (Moderate-Resolution Imaging Spectroradiometer) satellite image offers a large choice of opportunities for operational application in a large area. Study area is selected and located with the help of time series (June 1st week to October 2nd week MODIS Images) which provides maximum accuracy. A concept of assessing areas of potential cultivation of Soybean is suggested by means of GIS technique. Integrating MODIS data with GIS techniques and time series based assessment (June 1st week to October 2nd week) the actual areas of Soybean fields have been mapped and yield production have been predicted. The Soybean acreage area and production accuracy is 80.72% and there is a 7% increase (compared to 2006) in 2007. Introduction: United States, China, Brazil, Argentina and India are the five topmost producer of Soybean in the world. All the producers of soybean add up to a world total production of around 185 million tons. The annual production of soybean in India is around 7 million tons. Madhya Pradesh being the leading producing state of India contributes to around 75% of the total Indian production and is also called the soybean bowl of India. The cultivation of soybean in India is dependent on monsoon. Delayed monsoon and uneven rainfall, affects the acreage and yield of soybean leading to variation in the production value of Soybean. Since Soybean is a major agricultural commodity which is being exported from India, a prior knowledge of the expected acreage and production of soybean is very important for different planning purposes. The utilization of the predicted Crop Acreage & Production data is for both the Soybean Exports and also for the Government 1. Exporters require data on: o Produce of soybean that would be available for export o Base price that can be fixed based on supply and demand o Time by which the soybean may reach the Mandi 2. Government requires data on: o Total acreage of soybean in the state o Expected yield of soybean o Total expected production of soybean during the current year

o Tentative revenue collection Traditionally acquiring data about yield and acreage of crops is an extremely tedious job, including extensive travel and various interpolation methods based on the sample taken. Currently the agriculture department officials visit the village or tahsil where they inquire about crop acreage and expected yield. Based on these types of sampling the results are projected to acquire the acreage and yield information. This methodology, though prevalent from a long time is neither very accurate nor very scientific. On the other hand, it is predicted on the basic production pattern over the previous year. Yield depends on various factors, like climatic, physical etc, condition may or may not change over two successive years, however it is simple gambling that the climate condition of this year identical which is what is traditionally done for yield prediction. Apart from being cumbersome, counter cost effective and lengthy, these traditional methods are also too generalized and can't be fully relied upon. Alternatively, Remote sensing is a very important tool for study of soybean field. With the help of Remote sensing regional mapping is possible in a short time with much accuracy. Several studies have proved that the yield of soybean crop is directly related to the NDVI values of soybean calculated using optical image. The comparison of NDVI value of soybean for one year to another gives the quantitative value by which the yield value can differ from one year to another. In the present study an attempt has been made to calculate the yield of soybean crop for Kharif season, 2007, after comparison and correlation of NDVI and Yield values for the years 2005, and 2006. The point worth nothing is that here all of the cropped area is taken into account and not just sample sites. Since there was no spectral signature in form of GPS attribute available for historical years, ancillary crop acreage data collected from state government or other agencies, and general NDVI value ranges for each crop type, has been used for classification of different crop types. This study provides an alternative, accurate, fast and economic method of acreage estimation and yield prediction. In India Madhya Pradesh (M.P.) is the leading state in producing soybean followed by Maharashtra, Rajasthan and Uttar Pradesh. On average, Madhya Pradesh produces 66 percent of India's total soybean crop; Maharashtra 24% and Rajasthan 6%. Soybean grows well in warm and moist climate. The climatic requirements for soybean are almost the same as for maize. A temperature of 26.5 to 30°C appears to be the optimum for most of the varieties. Soil temperatures of 15.5°C or above favor rapid germination and vigorous seedling growth. The minimum temperature for effective growth is about 10°C. The time of planting is a very important consideration in soybean. In northern India soybean can be planted from third week of June to first fortnight of July. Latest research results from the Department of Agricultural Research and Education, ICAR, Government of India, have shown that planting soybean in the last week of June results in maximum yield and after 7 July causes reduction in seed yield. Well drained and fertile loam soils with a pH between 6.0 and 7.5 are most suitable for the cultivation of soybean.

The main objective of the present study is to estimate the acreage and yield of soybeans crop in some districts of M.P. in the Kharif season 2007. It provides statistical analysis in GIS environment and mapping of soybean crop. 11 districts are listed below, chosen for present study.

Fig-1 Location map of study area Bhopal, Lat 23°07’ to 23°54’ N, Long 77°12’ to 77°40’ E, Area 2,772 sq. km Dewas, Lat 20º17' to 23º20' N, Long 75º54' to 77º08' E, Area 7,020 sq. km Guna, Lat 24º53’ N to 25º6'55 N, Long 77º48' E to 78º16' E, Area 6,308 sq. km Indore, Lat 22°02’ to 23° 05’ N, Long 75°25 to 76°16’ E, Area 3,898 sq. km Raisen, Lat 22°47' to 23°33' N, Lang 77°21' to 78°49' E, Area 8,395 sq. km Rajgarh, Lat 23°27' to 24°17' N, Long 76°11' to 77°14' E, Area 6,154 sq. km Sagar, Lat 23°10’ to 24°27’ N, Long 78°4’ to 79°21’ E, Area 6,375 sq. km Sehore, Lat 22°31' to 23°40' N, Long 76°22' to 78°08' E, Area 6,578 sq. km Shajapur, Lat 23º06' to 24º19' N, Long 75º41' to 77º02' E, Area 6,196 sq.km Ujjain, Lat 23º18’ to 24019’ N, Long 75º45’ to 76058 E, Area 6,091 sq. km Vidisha, Lat 23°21' to 24°22' N, Long 77°15' to 78°18' E, Area 7,371 sq. km

Fig-2 Input Image in parts of Madhya Pradesh (MODIS Terra 250 m. Resolution, 7-2-1 Band Combination) 1June 2007 with district boundary

Fig-3 Input Image in parts of Madhya Pradesh (MODIS Terra 250 m. Resolution, 7-2-1 Band Combination) 1october 2007 with district boundary MODIS satellite data and topographical sheets were used for Remote Sensing and GIS analysis. Maps showing the administrative boundary of Soybean cultivated area. Methodology:

2005 Image

2006 Image

Geo- referencing

Geo- referencing

Administrative boundary overlaying

Administrative boundary overlaying

Sub- Setting Study Area

Sub- Setting Study Area

Classification

Classification

2007 Image

Geo- referencing

Administrative boundary overlaying

Sub- Setting Study Area

Classification

Calculate the Acreage Estimation

Flow Chart of Acreage Estimation The methodology essentially consisted of selection of the datasets, processing of the satellite data, incorporation of ground information, analysis of the satellite data, and generation of the output products. Images of the following dates are used, 1 June, 19 July, 23 August, 12 September and 1 October 2007, 2006, 2005 dates images. First June data is pre-cultivation data, while that of 19th July, 23rd August and 12th September show growing period and 1st October shows mature stage. Interesting part is that, 19th July, 23rd August and 12th September also show mixed crops in various growing stages, however in 1st October image only soybean is depicted as other crops have been harvested. We had images of soyabean growing areas in its mature stage for three successive years. We performed NDVI on all the october images (2007,2006,2005). We then compared the scene average NDVI values for individual districts over the years, to get the predicted yield of soybean. Hybrid classification techniques were used to show the spatial distribution of soybean in the different districts of the study area. This included classification of multi-temporal satellite images using unsupervised, supervised and time series based techniques. Acreage Estimation

District wise acreage estimation of Soybean was done by using the following mathematical calculationsArea in meter2 =Numbers of pixels of clusters * Resolution of the image Area in hectare= Area in meter2 / 10000 For MODIS - Area in meter2 =Numbers of pixels of clusters * 250*250 2005 Image

2006 Image

Masking Process (Removed Other Crops)

NDVI Value

Co-Relation of Production

Masking Process (Removed Other Crops)

NDVI Value

Co-Relation of Production

2007 Image

Masking Process (Removed Other Crops)

NDVI Value

Co-Relation of Production

Soybean Prediction in 2007 Flow Chart of Yield Estimation The following Flow Chart shows the general methodology adopted for the estimation of crop yield and prediction of production for the different districts of the study area. Crop yield was calculated by comparison of the NDVI** values of Soybean for the years 2005, 2006 and 2007. For this purpose appropriate cloud free satellite images during the heading stage of the soybean crop were selected for all the three years. These appropriate satellite images were then masked with the map showing the spatial distribution of soybean for the corresponding satellite images for the respective year. Further NDVI was calculated for all the three years. Statistical analysis was done between the NDVI values and Production values of Soybean for the years 2005 and 2006. The correlation factor was formulated between the NDVI and Crop yield. This correlation factor was used predict the Yield of Soybean for the year 2007, based on the NDVI value of Soybean for the same year.

The vintage and the NDVI value of soybean for the three years has been given in the table belowTable-1 Extract NDVI values 2005, 2006, 2007 in parts of Madhya Pradesh. YEAR 2005

DATE 12 September

NDVI 0.725

2006

15th September

0.731

2007

13th September

0.742

th

The production estimation was done using the following mathematical calculationsProduction= Estimated Acreage * Predicted Yield Result: The process indicated in methodology has been used to extract the Soybean area in parts of Madhya Pradesh. The estimated Soybean area of Madhya Pradesh (districts-wise) is given in the Se No DISTRICT NAME SOYBEAN CROP SATELLITE(MODIS) AREA IN ACREAGE AREA table PIXEL RESOLUTION SQUERE METER IN HAC(2007-2008)

1 2 3 4 5 6 7 8 9 10 11

Raisen Ujjain Bhopal Dewas Guna Indore Rajgarh Sagar Sehore Shajapur Vidisha

18776 67377 15063 45307 26888 34469 46711 49173 40850 49294 25070

(250*250) 250*250 250*250 250*250 250*250 250*250 250*250 250*250 250*250 250*250 250*250 250*250

(250*250/10000) 1173500000 4211062500 941437500 2831687500 1680500000 2154312500 2919437500 3073312500 2553125000 3080875000 1566875000

117350 421106 94143 283168 168050 215431 297943 307331 255312 308087 156687

below. On the image, green color is assigned to Soybean in the table here gives the pixels and calculated area in parts of Madhya Pradesh district wise. Table-2 Acreage of Soybean crop in parts of Madhya Pradesh

Kharif crop Soybean Acreage in ha. Satellite Based SOPA Acreage - SOPA Acreage S.No. Districts Acreage - 2007 SOPA Acreage - 2007 Difference 2006 2005 1 Ujjain 421,106 425,560 -4,454 412,900 406,300 2 Dewas 283,168 281,430 1,738 256,000 270,400 3 Shajapur 308,087 319,860 -11,773 303,000 304,200 4 Bhopal 94,143 85,230 8,913 82,500 85,500 5 Sehore 255,312 247,900 7,412 229,200 234,000 6 Raisen 117,350 95,200 22,150 80,500 63,400 7 Vidisha 156,687 144,750 11,937 143,600 90,000 8 Rajgarh 291,943 270,950 20,993 257,400 174,500 9 Indore 215,431 215,421 10 215,400 213,900 10 Sagar 307,331 202,000 105,331 178,200 95,200 11 Guna 168,050 164,000 4,050 132,000 117,200 years 2005 & 2006

Tabl e-3 Com paris on of acre age of year 2007 with acre age in the

The total acreage of soybean estimated using RS and GIS techniques in the different major districts of Madhya Pradesh is 26.5 in Lakh ha. Based on the satellite derived estimates of soybean acreage, Ujjain district has been adjudged the largest grower of Soybean during the current Kharif season with an estimated acreage of 421,106 ha. The other two major soybean growing districts, Shajapur (3, 08,087 ha) and Sagar (3,07,331 ha) have almost similar acreage of area under soybean. In Bhopal, although the acreage under soybean has increased as compared to last two years, but it is still the district with least acreage of soybean among the districts under study.

Fig-4 Resulted Classified Image Data of Soybean in Districts-wise Satellite based estimation of crop yield shows an average yield of 844 kg/ha for all the major soybean growing districts of Madhya Pradesh. Indore has shown a remarkable growth in the crop yield due to its improved farm management practices. This year Indore district has topped the list with an estimated yield of 964 Kg/ha. Ujjain and Sagar are the two other districts with an average yield of more than 900 kg/ha. The lowest yield of 730 kg/ha was recorded in the district of Sehore.

Table-4 Comparison of Satellite Estimated yield of year 2007 with that of field derived yield in the years in 2005 & 2006 Districts Ujjain Dewas Sahajahanpur Bhopal Sehore Raisen Vidisha Rajgarh Indore Sagar Guna AVERAGE YIELD

yield(kg/ha) 2005 920 910 930 650 770 850 830 850 780 790 850 830

Fig-5 Multi dates Modis images NDVI values of soybean crop in 2007

yield(kg/ha) 2006 930 830 890 730 720 810 670 870 950 890 870 832

yield(kg/ha) 2007 943 842 903 740 730 822 680 883 964 903 883 844

2007

13th Septem ber

2006

15th Septem ber

2005

12th Septem ber

0.742

0.731

0.725

0.715

0.72

0.725

0.73

0.735

0.74

0.745

Soyabin Yield Comparing 2005, 2006 & 2007

Fig-6 NDVI values soybean crop 2005, 2006 and 2007 in parts of Madhya Pradesh on the Bar Chart Total production of soybean in the 11 districts of Madhya Pradesh has been predicted to be 22.58 tonnes. Ujjain with acreage of 4.21 lakh ha and yield of 943 kg/ha is the largest producer of soybean in the state. The total production of soybean in Ujjain has been estimated to be 3.97 lakh tonnes. The lowest production of soybean has been estimated to be 0.69 lakh tonnes in Bhopal district. Area in Lakh ha and Production Lakh tonne Table-5 Total Soybean production in parts of M.P. Districts

Ujjain Dewas Sahjahanpur Bhopal Sehore Raisen Vidisha Rajgarh Indore Sagar Guna Total

Area Shown (2007)

Yield (2007)

Total Production

4.21 2.83 3.08 0.94 2.55 1.17 1.56 2.91 2.15 3.07 1.68 26.15

943 842 903 74 730 822 680 883 964 903 883 8627

3.97 2.38 2.78 0.69 1.86 0.96 1.06 2.56 2.07 2.77 1.48 22.58

Conclusion: The introduced methodology is promising. It enhanced the quantitative accuracy of estimation of acreage and production of soybean crop in parts of Madhya Pradesh 2007 for MODIS-based image classification. It ensured that the soybean cultivation areas were exactly extracted. The concept of GIS was employed to support spatial information management, and analysis of spatial data combined with related MODIS data was used to support the soybean area extraction process, using time series based classification for training samples allowed for accurate calculation of spectral parameters of land use classes and especially of areas of soybean

fields. GIS was a very helpful tool to pinpoint the location of potential areas of soybean fields. The classification accuracy of areas soybean cultivation by means of remote sensing techniques was thus significantly increased. The two 250 m bands of MODIS provided moderate spatial resolution and daily observation of the land surface. It had been verified that these data are well suited for mapping areas of soybean fields and for assessing growth parameters of soybean. Using the integration of GIS, MODIS-based image analysis enhances the quantitative and qualitative accuracy of classification of soybean fields. The methodology proved to be feasible, repeatable, and fast, since the GIS database and GPS were available. The range of relative error of classification was small. The methodology can be effectively used for the accurate classification of soybean fields based on daily available MODIS earth observation data. So Remote Sensing and GIS is the most important tools for study of Soybean crop, in this study we find out the Acreage Estimation, Yield and Production of Soybean crop in parts of Madhya Pradesh 2007 through remote sensing and GIS techniques. Acknowledgement: Authors are thankful to Remote Sensing and GIS Lab, MGCGV Chitrakoot for providing the lab facility for the present study. References: [1] R.S. DeFries, M.C. Hansen, J.R.G. Townshend, and R.S. Sohlberg, Global land cover classifications at 8 km spatial resolution: the use of training data derived from Land sat imagery in decision tree classifiers, International Journal of Remote Sensing, 19, 3141– 3168, 1998. [2]M.A. Friedl and C.E. Brodley, Decision tree classification of land cover from remotely sensed data, Remote Sensing of Environment, 61, 3, 399– 409, 1997. [3]R. Lawrence, A. Bunn, S. Powell, and M. Zambon, Classification of remotely sensed imagery using stochastic gradient boosting as a refinement of classification tree analysis, Remote Sensing of Environment, 90, 331–336, 2004. [4]R.S. DeFries and J.R.G. Townshend, NDVI-derived land cover classification at a global scale, International Journal of Remote Sensing, 15, 3567–3586, 1994. [5]P. Jonsson and L. Eklundh, TIMESAT––A program for analyzing time series of satellite sensor data, Computer & Geosciences, 30, 833– 845, 2004. [6]L.G. Ferraira, and A.R. Huete, Assessing the seasonal dynamics of the Brazilian Cerrado vegetation through the use of spectral vegetation indices, International Journal of Remote Sensing, 2510, 1837–1860, 2004. [7]J.F. Knight, R.L. Lunetta, J. Ediriwickrema, and S. Khorram, Regional scale land cover characterization using MODIS NDVI 250 m multitemporal imagery: a phenology-based approach, GIScience and Remote Sensing, 43(1), 1–23, 2006. [8]P. Jonsson and L. Eklundh, TIMESAT––A program for analyzing time series of satellite sensor data, Computer & Geosciences, 30, 833– 845, 2004. [9]P.C. Doraiswamy, B. Akhmedov, and A. Stern, Improved techniques for crop classification using MODIS imagery, Presentation at the International Geosciences and Remote Sensing Symposium, July 31– August 4, 2006, Denver, Colorado [CD].

[10]H. Zhang, M.A. Friedl, C.B. Schaaf, A.H. Strahler, J.C.F. Hodges, F. Gao, B.C. Reed, and A. Huete, Monitoring vegetation phenology using MODIS, Remtoe Sensing of Environment 84, 471–475, 2003.

Paper Reference No.: 178 Title of the paper: Estimation of Acreage & Crop Production through Remote Sensing & GIS Technique Name of the Presenter: Abhishek Kumar Maurya Author (s) Affiliation: Abhishek Kumar Maurya, Research Scholar, MGCGV Chitrakoot Satna M.P. Dr. Shashikant Tripathi, Senior Lecturer, MGCGV Chitrakoot Satna M.P. Sandeep Kumar Soni, Research Scholar, MGCGV Chitrakoot Satna M.P. Pradeep Kumar Soni, GIS Analyst, COWI India Pvt. Ltd. New Delhi. Mailing Address (s): Abhishek Kr Maurya s/o B.P.Maurya, Near Balaji Temple, Dhus ka maidan, Purani Bazar Karwi Chitrakoot U.P. 210205, India . Email Address: [email protected], [email protected], [email protected] Telephone number (s): +91-9369081090(Abhishek), +91-9795666553(Sandeep) Fax number (s): Not present. Author(s) Photograph:

Abhishek Kr. Maurya Sandeep Kr. Soni Brief Biography (100 words):

Pradeep Kr. Soni

Dear Sir, I am a research scholar in the subject Remote Sensing & GIS in MGCGV Chitrakoot Satna M.P. I also have Master Degree in the same subject. I have presented 3 research papers in different National Seminars and 1 research paper has been published in ezine weekly online publication of GIS Development. I have 2year working experience in relevant field and I have used many more GIS & Remote Sensing software and Satellite data.