search this site

These data products are being distributed free of charge. Recipients have a responsibility to:

1. Cite the following reference:

Frolking, Steve, J.B. Yeluripati, E. Douglas. New district-level maps of rice cropping in India: A foundation for scientific input into policy assessment. In press in Field Crops Research.

2. Acknowledge the University of New Hampshire, EOS-WEBSTER Earth Science Information Partner (ESIP) as the data distributor for this dataset.

In addition, it is recommended that users of these data contact Steve Frolking (steve.frolking@unh.edu ) to ensure proper data use and interpretation.

The Indian Rice Cropping Maps data collection can be ordered from the EOS-WEBSTER Search and Retrieve Tool

Summary:

We combined several district-level and state-level data sets of rice cropping in India to develop a single dataset of district level rice cropping systems for all of India in 1999-2000. The data set contains district-level areas for 34 different single-, double-, and triple-cropping combinations (e.g., Rice-Rice, Rice-Rice-Pulse, Rice-Pulse-Fibrecrop, Rice-Wheat, and Rice-Fallow). The dataset specifies cropping by season (e.g., Kharif and Rabi) and area in two water management systems (irrigated or rainfed) for each cropping system in each district. The total rice sown area is 44.9 million hectares (Mha), 91% in the Kharif (wet) season. Total rice land area was 41.6 Mha, with upland rice accounting for 13% of this area. Rainfed rice (including upland and deepwater) accounted for 44% of the total rice sown area, and Rice-Fallow for 38% of the total rice sown area. The total multiple cropping area with rice occupied 17.6 Mha, with dominant systems being Rice-Wheat (8.8 Mha), Rice-Pulse (3.2 Mha), Rice-Rice (2.2 Mha), and Rice-Oilseed (1.2 Mha, including Rice-Groundnut). We combined these maps with a simple, monthly time-step water balance model to estimate irrigation water demand for irrigated rice by district; total national demand was 200 km³ y ^-1. A complete national, district-level set of maps and data of rice cropping systems and water management will be useful as inputs to a range of studies on agricultural productivity, resource use, and environmental impacts of rice agriculture.

Variable Descriptions:

1. Crop rotations reported (all in hectares per district in 1999-2000)

1.1. Single cropping (one crop harvested in a year; n =5)

1.2. Double cropping (two crops harvested in a year; n = 26)

1.3. Triple cropping (three crops harvested in a year; n = 24)

1.4. key to crop rotation names:

r = rice
b = barley
f = fibrecrop
fal = fallow
g = groundnut
m = millet
o = oilseed
pot = potato
s = sorghum
sug = sugarcane
v = vegetable
w = wheat
upland = rice grown in upland cropping (not flooded paddy) management
deepwater = rice grown in highly flooded conditions (water depth > 1 m)

*rf = rainfed (all rice rotations are irrigated unless specified as rainfed at the end of their name, i.e., ‘…_rf'; irrigation data and specification only for rice, not for other crops)

In all cases, first season listed is Kharif season.

2. Irrigation water requirement for paddy rice

Irrigation water is an estimate of irrigation water demand by irrigated rice, in cubic kilometers per month. Actual irrigation water use in a particular year would depend on water availability, actual precipitation (values were calculated for a 1970-2000 climatology), infrastructure capacity to deliver water (pumps, energy supply, canal integrity, etc.)

Spatial Scale:

District-wide data (n = 512). District map from International Rice Research Institute (Huke and Huke 1997), represents political boundaries from roughly the mid-1990s. There have been some changes in political boundaries since then, including the formation of two new states.

Crop areas: Areas (hectares per district) are reported for each of 55 different rice cropping systems.
Irrigation water: Irrigation water requirements, for irrigated paddy rice only, are in cubic kilometers per month.

Data Format:

These data are available in ASCII Text format.

India Cropping Maps:

Deepwater Rice Area Double Rice Cropping Area Fraction of Rice Irrigated Fraction of Rice Rainfed India States and Regions Paddy Rice Annual Irrigation Water Requirements Rice-based Single Cropping Area (excludes uplands and deepwater) Rice-based Double Cropping Area Rice-based Triple Cropping Area Single Rice Cropping Area Triple Rice Cropping Area Total Rice Sown Area Upland Rice Area

View All India Crop Map Images (Recommended for Safari browsers)

Methodology Used to Construct India Rice Crop Rotations:

We used five primary data sources to develop district-level rice crop rotations for all of India (Table 1):

• Huke and Huke (1997) (hereafter H&H) provided district-level rice sown areas by water management: upland, deepwater, wet-season irrigated, dry-season irrigated, and rainfed rice, based on data from the early- to mid-1990s; total rice sown area for India was 42.8 Mha.

• The Directorate of Rice Development (DRD, 2004) provided district-level rice sown area in 1999-2000 (total of Kharif plus Rabi seasons); total rice sown area for India was 44.9 Mha.

• Yadav and Subba Rao (2001) (hereafter Y&SR) provided areas of the three primary crop rotations of most districts in most states of India, but no information on less dominant rotations. Altogether they provide district level areas for 105 different rice cropping systems (e.g., Rice-Rice-Rice, Rice-Wheat, Rice-Fallow), as well as numerous non-rice rotations (e.g, Wheat-Maize) for 400 districts in 17 states. T otal reported sown area of rice was 27.6 Mha. We first simplified the list of crop rotations by combining similar crops into single classes (e.g., all grams and other pulses into a single ‘pulse' class), so the final crop types were rice, pulse, wheat, potato, oilseed, fibrecrop, groundnut, millet, vegetable, barley, sorghum, sugarcane, and fallow. There were no data for some districts within states for which Y&SR reported data. We filled these gaps by assigning the district's total rice area from DRD data across all Y&SR crop rotations based on the state average proportions in each rotation for districts that had reported values. The total rice area allocated in this manner was 1.75 Mha in 14 districts. Finally, we aggregated together similar rotations in which one representative had much less area than the other. For example, Rice-Rice-Pulse had 40,761 ha in Y&SR, while Rice-Pulse-Rice had 904 ha; we combined this into 41,665 ha of Rice-Rice-Pulse. This reduced the total number of Y&SR rotations to 34 (Table 2). In all cases, we have assumed that the first rotation listed is Kharif-season, although this is not specified in Y&SR.

• There were no Y&SR data for the complete states of Sikkim, Arunachal Pradesh, Nagaland, Mizoram, Meghalaya, Tripura, Manipur, Goa, the Union Territories, and the Andaman & Nicobar Islands. In the DRD database, these states contain a rice area of 0.9 Mha. We supplemented the Y&SR dataset with estimates of areas of three dominant crop rotations for these remaining states. For states in northeastern India ( Sikkim, Arunachal Pradesh, Nagaland, Mizoram, Meghalaya, Tripura, and Manipur) , we used season-wise crop area data from the North Eastern Region Databank of India (NERDB, 2005) to estimate crop rotations. We first scaled district-wise rice area from H&H to the DRD district-wise total areas. We then identified the dominant three crop rotations based on season-wise crop data from the NERDB database. We sorted seasonal crop data by area, and created double-cropping rotations based on the dominant crops in each season. The dominant rotation would be a Kharif-Rabi rotation of the two seasonal dominant crops, with an area equal to the smaller of the two seasonal dominant crops. These areas were then subtracted from the seasonal crop data, and the second dominant double crop rotation was determined in the same manner, and then the third. For Goa and the Union Territories (Dadra & N. Haveli, Diman, Diu, Karaikal, Mahe, Pondicherry, and Yanam), we used the same major crop rotations as their neighboring, and much larger states, scaling areas to the same proportion of total rice sown are. For the Andaman & Nicobar Islands we assumed that all rice was in a rice-fallow rotation, as we had no independent data. This complete dataset is hereafter called Y&SR*.

• The Regional Data Exchange System of the FAO Regional Office for As ia and the Pacific (FAO-RAP, 2005) reported two additional state-level data sets for 1999-2000: Rabi-season rice area and fraction of total rice area that is irrigated. The data (hereafter DAC-MoA) are from the Department of Agriculture & Co-operation, Ministry of Agriculture, Government of India.

We based our analysis on the district-level data of Huke and Huke (1997), reporting sown area by water management for the early- to mid-1990s. We updated the H&H data set in three steps to year 2000 areas, and then partitioned this data into rainfed and irrigated crop rotations in two additional steps (Table 2).

Updating district-wise total sown area to year 2000 values

First we scaled district total rice area from the H&H value, generally representing the early 1990s, to the 1999-2000 district total rice area reported in DRD. The H&H area in each water-management class, , was multiplied by the ratio of the DRD district-level total rice area, DRD i , divided by the H&H district total rice area.

 

(1)

where the subscript i identifies a district, and the subscript j identifies a water-management (e.g., rainfed, deepwater, etc.). Overall, this increased the total rice area of India by 4.9% (2.1 Mha), though for individual districts the increase could be larger or smaller, or could be a decrease.

For about 10 districts DRD did not report values, so we used the H&H values without scaling.

Updating district-wise Rabi-season sown area to year 2000 values

Second, we adjusted the district-level H&H area in Rabi-season rice (‘dry-season irrigated' in H&H) so that state-level Rabi-season totals matched the DAC-MoA values. To do this, for each state, we adjusted district-level Rabi-season rice area as follows

(2)

where   is the state total Rabi-season rice area reported in the FAO database, and the summation is over all districts, i , within the state. During this analysis, we concluded that H&H mis-reported irrigated areas for the state of Tamil Nadu, swapping dry-season irrigated and wet-season irrigated. H&H reported 1990 areas of 1.50 Mha of Rabi (‘dry-season irrigated') and 0.33 Mha of Kharif (‘wet-season irrigated') for Tamil Nadu, while the FAO database reported 1999-2000 areas 0.25 Mha of Rabi-season rice and 1.91 Mha of Kharif-season rice. We therefore swapped the H&H dry-season irrigated and wet-season irrigated area values for all districts in Tamil Nadu before applying Equation 2. For all states, we assumed that district-level gains (losses) in Rabi-season rice area were offset by district-level losses (gains) in rainfed rice area, and not by changes in district-level upland or deepwater rice. Nationally, this added 0.78 Mha of Rabi-season area, and reduced the rainfed area by 0.78 Mha.

Updating district-wise Kharif-season irrigated sown area to year 2000 values

Finally, since rice irrigated area in India increased in the 1990s from 19.4 to 24.3 Mha (Chanda et al., 2003), we adjusted the district-level Kharif-season irrigated rice (‘wet-season irrigated' in H&H) so that the total state-level irrigated rice (wet-season + dry-season) matched the total Ministry of Agriculture values reported in Chanda et al., (2003). We assumed that district-level gains (losses) in Kharif-season irrigated rice area were offset by district-level losses (gains) in rainfed rice area, and not by changes in district-level upland or deepwater rice.

(3)

 

The fraction of rice that is irrigated, f , is given for each dataset by

 

(4)

We limited changes in area so that neither final area was negative. We did not apply Equation 3 to the state of Assam because the data reported in Chanda et al., (2003) were from 1953-54. We did not apply Equation 3 to the Union Territories (districts of Dadra & N. Haveli, Diman, Diu, Karaikal, Mahe, Pondicherry, and Yanam ) because no data were reported in Chanda et al. (2003). For some states this decreased Kharif-season irrigated area (i.e., DAC-MoA irrigated fraction was less than H&H irrigated fraction), while for other states this increased Kharif-season irrigated area. The biggest increases were in the states of Uttar Pradesh (0.8 Mha), West Bengal (0.7 Mha), Madhya Pradesh (0.3 Mha), and Bihar (0.2 Mha). Nationally, this added 2.2 Mha of irrigated area, and reduced the rainfed area by 2.2 Mha.

Partitioning sown area into crop rotations

To disaggregate the district-level rice areas into different crop rotations, we used the dataset of Yadav and Subba Rao (2001), supplemented by NERDB (2005). We scaled the Y&SR* district-level rice cropping system rotation areas as follows. First, if the total district-level rice area for the three crop rotations reported in Y&SR* was greater than the DRD district-level total, the Y&SR* values were all reduced proportionally so that their total matched the DRD total; otherwise they were kept unchanged. Second, if the total district-level Rabi-season rice area for the three crop rotations reported in Y&SR* was greater than the Rabi-season total in the updated database, these rotations were scaled to match that total, otherwise they were unchanged.

We then assigned rice areas to upland, deepwater, and all Y&SR* rotations based on the scaled values discussed above. The remaining district-level rice area, which might be zero, was assigned to three rotations: rice-rice, rice-fallow, and/or fallow-rice rotations. First, an area equal to the minimum of the remaining areas of Rabi-season and Kharif-season rice was added to the rice-rice rotation (note that this minimum could be zero if all of the rice from one season had already been assigned to Y&SR* rotations), and the remaining areas of Rabi-season and Kharif-season rice were reduced by this amount, bringing the smaller value to zero. Then the remaining area in Kharif-season rice, which might be zero, was added to the rice-fallow rotation and the remaining area in Rabi -season rice, which might be zero, was added to the fallow-rice rotation.

Partitioning crop rotations into rainfed and irrigated areas

Most districts had more total rice sown area than total irrigated rice area, so the final step was to partition some rotations into irrigated and rainfed subsets. We assumed that all upland and deepwater rice was rainfed. We assumed that all rice crops in rotations with Rabi-season rice were irrigated (i.e., Rice-Rice-Rice, Rice-Rice, the 8 Rice-Rice-Other and Rice-Other-Rice rotations, the one Other-Rice-Other rotation, the two Other-Rice rotations, and Fallow-Rice). Depending on areas of these rotations and of Kharif-season irrigated rice, this allocation accounted for none, some, or all of the Kharif-season irrigated rice area in a given district; any remaining Kharif-season irrigated area was allocated to the remaining Kharif-season-only rotations in that district, with the following prioritization: first, to all 7 triple-crop rotations (Rice-Other-Other), proportional to their areas in the district; second, to all 11 double-crop rotations (Rice-Other), proportional to their areas in the district; and finally, any remaining Kharif irrigated area was allocated to Rice-Fallow. Thus, these 19 Kharif-season-only rice crop rotations could have irrigated and rainfed areas, while all other rotations were irrigated-only, except upland and deepwater, which were rainfed-only.

Determining irrigation water requirements

We used the district-season, season-wise area of irrigated rice and a simple monthly water balance model to estimate monthly, district-wise paddy rice irrigation water requirements for all of India. Irrigation water demand, IRR , was calculated on a monthly basis for irrigated rice as follows. The change in soil water storage ( DSW ) is given by

(5)

where P is precipitation, PET is potential evapotranspiration, and DR is drainage or percolation losses, all in mm month -1 . PET was estimated using the Shuttleworth and Wallace (1985) modification of the Penman-Monteith PET function, a physically-based method recommended by the Food and Agricultural Organization of the United Nations. Monthly precipitation data was based on the 1950-1995 climatology of the University of East Anglia Climate Research Unit (New et al., 1998); this gridded product was linearly interpolated to the center point of each district (Douglas et al., 2005-in review). Eq. 5 can then be solved for irrigation water demand by district and by month as

(6)

Percolation losses are primarily a function of soil texture. Guerra et al. (1998) report percolation rates of 1-5 mm d -1 for puddled clay and 24-29 mm d -1 for sandy loam or loamy sand. Kawaguchi and Kyuma (1977) used a mean value of 100 mm mo -1 for a pan-tropical Asia analysis. FAO (2004) report generic low and high percolation losses for paddy rice of 200 and 700 mm crop -1 . We determined the value for each district based on its ratio of average sand content to average clay content; soil texture properties were taken from a gridded (1° x 1°) global soil texture dataset (Webb et al., 2000). The highest district-level sand:clay ratio across India (8.9) was given a percolation rate of 600 mm mo -1 , the lowest sand:clay ratio across India (0.1) was given a percolation rate of 30 mm mo -1 , and all other districts were linearly interpolated between these two values, based on sand:clay ratios. The mean percolation rate was 142 mm mo -1 ( n = 439).

In planting months, the paddy soil is wetted from field capacity to saturation plus flooding; FAO (2004) provides a range of 150-250 mm water for land preparation. We approximated DSW as +250 mm for the district with the highest sand:clay ratio, +150 mm for the district with the lowest sand:clay ratio, and linearly interpolated between these values for all other districts. In the harvest month, soils are drained and allowed to dry, so we approximated DSW in those months as -150 mm to -250 mm, based on the water required for preparation . During rice growth months, the soil was maintained in a flooded state, and DSW was zero.

A generalized cropping season was used: Kharif-season rice – planting in July and harvest in November; Rabi-season rice – planting in January and harvest in May; double rice – Kharif and Rabi; triple rice – planting in January, May, and September, harvest in April, August, and December (Chanda et al. 2003). District-wide, monthly paddy rice irrigation water demand was then calculated as the product of IRR and the area of irrigated rice growing in each month. During fallow months and months with non-rice cropping, irrigation water demand was assumed to be zero (i.e., we calculated demand only due to rice cropping). Irrigation water demand was then aggregated spatially from district to state, regional, and national totals, and temporally from monthly to annual totals.

References cited:

Chanda, T.K., Sati, K., Robertson, C., 2003. Fertilizer Statistics . Fertilizer Association of India, New Delhi, 440 pp.

DRD ( Directorate of Rice Development ). 2004. Rice in India – A Handbook of Statistics ; on-line data ( http://dacnet.nic.in/rice/ ): District-wise Area, Production and Yield of Rice across the States during 1990-2000 ; Directorate of Rice Development, Ministry of Agriculture, Dept. of Agriculture & Co- Operation, Govt. of India; (visited 10-11 August 2004).

FAO. 2004. Rice Fact Sheet - Rice and Water . (www.fao.org/rice2004/en/factsheets.htm)

FAO-RAP. 2005. On-line database of the FAO Regional Office for As ia and the Pacific ( www.faorap-apcas.org/india/index.htm ), visited March 2005. (NOTE: see also FAO's new AGROMAPS web site: http://www.fao.org/landandwater/agll/agromaps/interactive/index.jsp ).

Guerra, L.C., Bhuiyan, S.I., Tuong, T.P., Barker, R., 1998. Producing More Rice with Less Water from Irrigated Systems, SWIM Paper 5, International Water Management Institute, Colombo, Sri Lanka, 33 pp.

Huke, R.E., Huke, E.H., 1997. Rice Area by Type of Culture: South, Southeast, and East Asia . IRRI, Los Baños, Philippines, 59 pp.

Kawaguchi, K., Kyuma, K., 1977. Paddy Soils in Tropical Asia: Their Material Nature and Fertility , Univ. Press of Hawaii, Honolulu, 258 pp.

NERDB. 2005. on-line database of the North Eastern Region Databank of India (http://databank.nedfi.com ); visited March 2005.

New, M., Hulme, M., Jones, P., 1998. Representing twentieth century space-time climate variability. Part II: Development of 1901-1996 monthly grids. J. Climate 13: 2217-2238.

Shuttleworth, W.J., Wallace, J.S., 1985. Evaporation from sparse crops: an energy combination theory, Quarterly J. R. Meteorol. Soc ., 111:839-855.

Webb, R.W., Rosenzweig, C.E., Levine, E.R., 2000. Global Soil Texture and Derived Water-Holding Capacities. Data set. Available on-line [http://www.daac.ornl.gov] from Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A.

Yadav, R.L., Subba Rao, A.V.M., 2001. Atlas of Cropping Systems in India . PDCSR Bulletin No. 2001-2, Project Directorate for Cropping Systems Research, Modipam, Meerut, India; 96 pp.

Table of datasets used to generate the new district level maps

Table 1 lists the datasets used and their characteristics. Table 2 provides an overview of the methodology used to assemble the new data set from the 6 original data sets.

Table 1 – Datasets used to generate the new district-level maps of rice cropping in India.

¹ Andhra Pradesh, Assam, Bihar, Gujarat, Haryana, Himachal Pradesh, Jammu & Kashmir, Karnataka, Kerala, Madhya Pradesh, Maharashtra, Orissa, Punjab, Rajasthan, Tamil Nadu, Uttar Pradesh, West Bengal.

² Sikkim, Arunachal Pradesh, Nagaland, Mizoram, Meghalaya, Tripura, and Manipur.

Data Sources:

DRD (Directorate of Rice Development). 2004. Rice in India – A Handbook of Statistics ; on-line data
(http://dacnet.nic.in/rice/): District-wise Area, Production and Yield of Rice across the States during 1990-2000 ; Directorate of Rice Development, Ministry of Agriculture, Dept. of Agriculture & Co- Operation, Govt. of India; (visited 10-11 August 2004).

FAO-RAP. 2005. On-line database of the FAO Regional Office for As ia and the Pacific (www.faorap-apcas.org/india/index.htm), visited March 2005. (NOTE: see also FAO's new AGROMAPS web site: http://www.fao.org/landandwater/agll/agromaps/interactive/index.jsp).

Huke, R.E., Huke, E.H., 1997. Rice Area by Type of Culture: South, Southeast, and East Asia . IRRI, Los Baños, Philippines, 59 pp.

NERDB. 2005. on-line database of the North Eastern Region Databank of India (http://databank.nedfi.com); visited March 2005.

Yadav, R.L., Subba Rao, A.V.M., 2001. Atlas of Cropping Systems in India . PDCSR Bulletin No. 2001-2, Project Directorate for Cropping Systems Research, Modipam, Meerut, India; 96 pp.

Table 2 – Procedure for combining datasets for district-wise maps of rice cropping area.

¹ discussed in Frolking et al. (2006) in section noted in parentheses.

²  H&H: district-wise sown area by water management (Huke and Huke, 1997); DRD: district-wise sown area (Directorate of Rice Development, 2004); DAC-MoA-1: state-wise Rabi-season sown area (FAO-RAP 2005); DAC-MoA-2: state-wise total irrigated sown area (FAO-RAP 2005); YSR*: district-wise land area by crop rotation (Yadav and Subba Rao, 2001; NERDB, 2005).

³  IW – irrigated wet season; ID – irrigated dry season; RF – rainfed; UP – upland; DW – deepwater.

Acknowledgement:

This research was supported by the National Aeronautics and Space Administration (NASA), under the Terrestrial Ecology Program (NAG5-12838) and EOS-IDS program (NAG5-10135).

Citation:

Frolking, Steve, J.B. Yeluripati, E. Douglas. New district-level maps of rice cropping in India: A foundation for scientific input into policy assessment. In press in Field Crops Research.

Data Providers:

Dr. Steve Frolking, Complex Systems Research Center, Institute for the Study of Earth, Oceans, and Space, Morse Hall, University of New Hampshire, Durham, New Hampshire, USA. Ph: 603.862.1792, Fax: 603.862.0188, Email: steve.frolking@unh.edu.

Latest Data Update:

4/07/2006

Last Doc. Updated:

5/19/2006

Doc. Updated By:

Steve Frolking

	search this site