Rice has become a staple food source in Indonesia, as evidenced by statistical data from the amount of rice consumption in 2018 reached around 29.57 Million Tons. Based on these data the determination of the paddy field area is an important aspect in determining the sustainability of food security. According to data from (Kementerian Pertanian, 2018) the area of paddy fields continues to change every year, in 2016 the area of paddy is 8,187 million hectares while in 2017 is 8,162 million hectares due to the dynamics of paddy plant growth tends to occur in a short time, so that monitoring of paddy is needed in near-real-time.

One of the method that is used in determining whether a land belongs to the paddy field or not is to use Remote Sensing approach. (Dong, et al., 2016) has mapped the paddy field area in Northeast Asia with the Remote Sensing approach using Landsat 8 Satellite Imagery. Paddy field mapping has also been carried out by (Cai, et al., 2019) using the time series Sentinel l-2 image to obtain paddy field map in the Dongting Lake area in Hunan Province, China and field survey data that used to validate the accuracy of paddy field map.

Every object on earth has a different spectral reflectance, even in one variety of paddy can have different wavelengths that can be influenced by the variety, age, density, to the level of health / greenness of the plant. Spectral Signature identification is a method of Remote Sensing. According to (Vaesen, et al., 2001) there are several approaches in identifying Spectral Signature such as RVI, NDVI, PVI, WDVI and LAI. Based on the approach it is expected that using Spectral Signature can classify paddy to the level of species, age, and health.

Monitoring and forecast local crop production are critical steps in addressing food security problems at a global scale. The combined effects of a changing climate, growing population, soil loss, as well as the natural variability of weather, require methods that provide a timely and accurate assessment of crop growth and production (Huang, et al., 2019). Crop growth and production model are urgently needed to dynamically simulate crop phenology, leaf area index, biomass, water use and grain yield formation in response to variations in genotype, environment and management, as well as their interactions (de Wit, et al., 2015).

A DSS is an information system that supports a user in choosing a consistent response for a particular problem in a reduced time frame (Hamouda, 2011). DSS are computer-based systems, built in order to solve multi-scenario problems by analyzing the feasibility of each scenario in a short time in order to provide a near optimum solution among them. A DSS may also be applicable for multiple problems and the possible solutions may or may not integrate aspects of sustainable development (Mannina, et al., 2019).

Precision agriculture is one approach that can be used in overcoming problems in agriculture. The purpose of precision farming itself is to minimize (spending costs, 7 time, and the spread of pests and diseases), optimize the use of resources effectively and efficiently, and maximize the harvest from agricultural land. As stated by (Zhang & Kovacs, 202) precision agriculture (PA) is the application of geospatial techniques and sensors (e.g., geographic information systems, remote sensing, GPS) to identify variations in the field and to deal with them using alternative strategies. According on that statement, the need for geospatial techniques that is integrated with information systems will be used as a tool to deliver information from government level to farmer group level.

The advance of technological development in agriculture is not impossible to develop an agricultural information system that can help various problems that occur in agricultural land, currently has developed several information systems about rice farming such as IPB Digitani (IPB University), SIMOTANDI (Rice Plantation Monitoring Information System ) from the Ministry of Agriculture, KATAM (Integrated Planting Calendar) which contains information related to the planting calendar developed by the BALINGBANGTAN (Agricultural Research and Development Agency) (Balitbangtan, 2015). This research is expected to build an information system that can be a tool for its users in terms of overcoming problems that arise in paddy agriculture in form of Web-GIS and Mobile App.

The general objective of this research is to develop spatial decision support system for village based national rice production (SIPANAS, Sistem Informasi Manajemen Padi Nasional). The specific objectives of this study are:

1. Identifying Paddy plantation to the level of the variety, age and health level based on the spectral signature reflectance of the paddy (1st year),

2. Developing a baseline rice growth and development model (1st, 2nd year),

3. Building a spatial-explicit dynamic model, in order to get (e.g. growth, yield, water balance, nutrient status and pest and disease) (1st, 2nd year),

4. Building a model to determine DSS for precision crop and field management (e.g. crop and field management, fertilizer management, water management, pest and disease management, CSA (Climate Smart Agriculture), adaptation/mitigation option and GHG (Greenhouse Gas) inventory) (2nd year).

5. Developing an information system as a means of delivering information for each stakeholder from government level to farmer groups level (3rd year).  

CONCLUSION

Some conclusions based on data that have been processed and analyzed are as follows:

1. The orientation of the experimental plot or plot in the research area duplicates the pixel arrangement in the Sentinel-2 image.

2. Weather conditions during the study were generally sunny or slightly cloudy and only rained 13 times, with a total rainfall of 78.8 mm. Even so, plants do not experience water shortages because there is a good supply of water from irrigation.

3. The development of the three rice varieties shows that Inpari-32 has a longer harvest life than IR-69 and Pandan Wangi varieties.

4. The dynamics of LAI are generally uniform, experiencing an increase in LAI until it reaches a maximum at the maximum vegetative growth stage, and then decreases towards harvest time. LAI characteristics need to be further analyzed to see their response to the treatment of fertilizer doses, varieties, and planting techniques

5. NDVI graphic display from Sentinel-2 image shows that: a. NDVI gave a visually more significant response to the fertilizer dosage treatment. A higher fertilizer dosage gives a higher NDVI value, especially in the early stages to the maximum vegetative stage. b. The difference in cropping patterns did not give NDVI responses which were visually significantly different. c. Different varieties did not respond to NDVI which was visually significantly different.

6. RGB drone images can produce differences in vegetation indices in response to fertilizer treatment, planting techniques, and varieties. Visually, the vegetation index was seen to be higher in the treatment of higher fertilizer doses, on the technique of planting rows of tiles, and in the Pandan Wangi variety.

7. Further processing is required to produce an empirical formula that represents the spatio-temporal relationship between signature reflectance and plant characteristics, plant and soil macro nutrient status and environmental conditions as a baseline for a spatial-explicit dynamic model of rice plants.