Spatial analysis of the relation of rice land area and rice production using remote sensing imagery

Indonesia is the country with the largest rice producer. One of the provinces that is a national rice granary is West Java Province, specifically Cianjur Regency which is the third largest contributor of rice products in West Java. In determining the area of rice fields, knowledge of spatial information is very necessary, processing can be done using remote sensing. Remote sensing is able to detect land use and land cover changes. The data needed is Landsat 8 imagery in 2013 and Landsat 8 imagery in 2020. This spatial analysis of land change passed an accuracy test using the Kappa Coefficient method. The results of the analysis carried out over a period of 7 years showed that there was an increase in the area of rice fields amounting to 16671.2 Ha. From this relationship, the 2013 image calculation gets a production value of 801,622 tons or 8.02 tons per hectare, and the 2020 image gets a production value of 1,108,388 tons or 7.89 tons per hectare.


Introduction
Agriculture is the activity of utilizing biological resources carried out by humans to produce food.The agricultural sector is a sector that continues to have an important role in the structure of the national economy.The role of agriculture in economic development is only seen as passive and merely a supporting element [1].In fact, the process of economic development is one of the roles of the agricultural sector [2].If a region wants smooth and sustainable development, then the region must start from the countryside in general, and the agricultural sector in particular [1].
In the context of sustainable agricultural development, land is the main resource in agricultural businesses, especially in conditions where most of the business sectors being developed still depend on agricultural patterns.Rice fields are agricultural land that is plotted and bordered by embankments.Rice fields play an important role in providing rice needs.Rice is a staple food for more than half of the world's population, making it very important for food security [3].94% of rice production in Indonesia is produced from wetland rice farming, the rest is produced from dryland rice farming.With an area of only 7% of Indonesia's land area, Java Island is still the largest contributor to rice production, reaching 53% of total national rice production [4].
West Java Province as one of the national rice granaries over the last 30 years, on average has contributed around 22% of national rice production and more than 95% of rice production is produced from rice fields, the rest is produced from dry land [5].Cianjur Regency is one of the rice barns in West Java that needs to be maintained and is the third largest contributor to rice production in West Java [6].
The availability of geospatial data and information is very necessary in various development activities.Geospatial data is data about geographic location, dimensions or sizes, and/or characteristics of natural and/or man-made objects that are under, on, or above the surface of the earth, while Geospatial Information is data that has been processed so that it can be used as a tool. in policy formulation, decision making, and/or implementation of activities related to terrestrial space [7].
The importance of Geospatial Information in mapping is to avoid mistakes, errors and overlapping information in making a map [13].Technology that can provide geospatial data and information quickly is remote sensing technology.Remote sensing is a way of depicting the condition of a particular area through a sensing device or sensor installed on a vehicle [8].Remote sensing has been recognized as an effective tool in detecting land use and land cover changes.
In this paper was examined the analysis of the relationship between rice field area and rice production in Cianjur Regency by utilizing Landsat 8 imagery for 2013 and 2020, so that changes in rice fields can be seen during that time.The object area of the research is Cianjur Regency, West Java.(Figure 1.)

Image Classification
In this research, Maximum Likelihood guided classification is used because this classification method is very effective when applied with very accurate data samples or training areas, and is also one of the most commonly used classification methods.According to [9], the algorithm is the maximum likelihood algorithm which is statistically the most well-established.If other algorithms are based on measuring the distance between the coordinates of the sample cluster and the coordinates of the candidate pixels (which will be classified or labeled), then the maximum likelihood algorithm uses basic probability calculations.
At this stage, the main thing is the accuracy of manual image interpretation.In this research, the author created 6 land cover classes that will be classified, because the number of land cover classes is adjusted to the existing land cover in Cianjur Regency.The results of the 6 cover classification classes are: 1) Water, 2) Forest, 3) Plantations, 4) Vacant Land, 5) Rice Fields, 6) Settlements.The following are the results of the classification of satellite images for 2013 and 2020 that have been carried out.Water usually shows with irregular flow patter and elongated shape.

Accuracy Test
Image classification results can be said to still be tentative if the accuracy testing process has not been carried out.Sometimes in a classification process, good or bad classification results depend on the accuracy of the interpretation technique used by the interpreter.The accuracy test in one flow of the image classification process aims to determine the level of truth of the classification model that has been created [10].
Total accuracy describes the overall truth value of the correct appearance of objects on the classification map with field data.From the matrix table above, the overall average value is calculated using the following formula: Apart from using the classification test above, there is one more parameter that is used to determine the level of truth of the classification results, namely the Kappa Coefficient method.The value range for the Kappa coefficient is from 0 -1 [3]. the Kappa coefficient accuracy level categories are: The classification test carried out in this study used the error matrix method.The following are the results of the image accuracy tests carried out.From the results of accuracy tests carried out on Landsat 8 images in 2013 and 2020, each image had an accuracy of 99.38% and 98.15%.while the kappa coefficients are 0.98% and 0.91% respectively.These results show that the two image classification results carried out are good quality.

Land Use in Cianjur Regency
The results of land analysis in Cianjur Regency showed that the land cover that occurred in each land cover from 2013 to 2020 experienced increases and decreases in area.The reduction in area occurred in parts of water bodies, forests, plantations, empty land which occurred within a period of seven years.Meanwhile, additional area occurred in rice fields and built-up land [14].For more details, see the information below.This rice field analysis was carried out using the overlay method in ArcGIS software.The Overlay method is a method of overlapping 2 or more maps to find out the desired spatial information.The use of rice fields has increased from 2013 to 141381.5 ha and 158052.6 ha in 2020.So in a period of 7 years From the research carried out, it was found that the area of rice fields was used for rice production data in Cianjur Regency from the Department of Agriculture, so a calculation process could be carried out to find out the relationship between the area of rice fields and rice production in Cianjur Regency.The formula of relation Rice Field Areas with Rice Production: The following is a comparison of rice fields and rice production in Cianjur Regency in 2013 and 2020.Based on the results of the calculation of the area of rice fields and rice production that has been carried out, it can be seen the relationship between the area of rice fields and rice production in Cianjur Regency.In this research, the results were obtained: • Calculation results in the 2013 image show a rice production value of 801,622 tons with an average of 8.02 (tons per hectare) of rice produced.• Calculation results in 2020 show a rice production value of 1,108,388 tons, with an average rice production of 7.89 (tons per hectare).

Figure 1 .
Figure 1.Study Map Area

Figure 2 .
Figure 2. Flow diagram of Research

Figure 2 .
Result of Accuracy Test from Satellite Image (a) 2013 and (b) 2020

3. 3 Figure 3 .
NDVI Transformation In this process, satellite images that have been classified and have gone through the accuracy testing stage will undergo NDVI transformation.The NDVI transformation was carried out to determine the vegetation index pattern for rice fields in Cianjur Regency in 2013 and 2020.So it can be seen changes in the range of vegetation index values for rice fields in 2013 and 2020.NDVI values are basic parameters derived from optical remote sensing data such as Landsat 8 satellite images which are used to detect the greenness of vegetation.The distribution of vegetation density values obtained based on images from 2013 and 2020 has different values.The following are the results of analyzing NDVI values on two different images.Result of NDVI values in Cianjur Regency Year (a) 2013 and (b) 2020

Table 2 .
Image Result of Visual Interpretation

Table 3 .
Kappa coefficient categories

Table 4 .
Accuracy Test Image Year 2013 and Year 2020

Table 5 .
Land Use Area in Cianjur Regency Figure 5. Land Use Map in Cianjur Regency 2013

Table 6 .
Area of Ricefields and Rice Production rice fields in Cianjur Regency increased by 16671.2ha, the increase in area was due to the transition of other areas to use of rice fields.

Table 7 .
Relation of Rice Field Areas with Rice Production in Cianjur RegencyThe results the analysis carried out using supervised classification, in a period of 7 years from 2013 to 2020, it was found that the area of rice fields in Cianjur Regency increased by 16671.2Ha.This happens because there is a very high shift in use of other land to rice fields.Cianjur Regency consists of 32 sub-districts spread out, with the area of rice fields experiencing an increase in the area of rice fields in 16 sub-districts, and 16 sub-districts experiencing a decrease.The sub-district that experienced the largest increase was in Naringgul District with an area of 9195.87Ha, while the largest reduction was in Karang Tengah District with 6213.35Ha.The largest area of rice fields in Cianjur Regency in 2013 was in Takokak District with an area of 9246 Ha of rice fields, while in 2020 the largest area of rice fields was in Naringgul District with an area of rice fields of 13367 Ha.Meanwhile, in 2013 the smallest area of rice fields was in Pacet District, 778 Ha, and in 2020, the smallest area was in Cianjur District with an area of rice fields of 565 Ha.