本郷 千春

An efficient assessment of rice varieties in tropical regions is crucial for selecting cultivars suited to unique environmental conditions. This study explores machine learning algorithms that leverage multispectral sensor data from UAVs to evaluate rice varieties. It focuses on three paddy rice types at different ages (six, nine, and twelve weeks after planting), analyzing data from four spectral bands and vegetation indices using various algorithms for classification. The results show that the neural network (NN) algorithm is superior, achieving an area under the curve value of 0.804. The twelfth week post-planting yielded the most accurate results, with green reflectance the dominant predictor, surpassing the traditional vegetation indices. This study demonstrates the rapid and effective classification of rice varieties using UAV-based multispectral sensors and NN algorithms to enhance agricultural practices and global food security.

Red crown rot (RCR) is a soil-borne disease that damages soybean growth and decreases yield. Infected plants show earlier defoliation and pencil-like roots, sometimes resulting in mortality. This disease became common relatively recently, and information about its field-scale appearance is insufficient. Insufficient data is a major constraint when planning countermeasures. In this study, unmanned aerial vehicle (UAV)-acquired images were used to visualize the spatial and time series variation in the area damaged by RCR in the same farmer fields in 2018 and 2020. Field investigation showed that RCR severely damaged soybean production. The reductions of yield were estimated at 17.5% and 12.7% in 2018 and 2020, respectively. The visualized damage clarified the difference in the increasing rate and patterns of RCR between the 2 years. In 2018, the damaged area expanded along the planting row to the whole field, but in 2020, the expansion along the planting row was not great, and half of the fields remained sparsely damage. This difference implies that various factors are associated with damage occurrence and pathogen distribution. The method applied in this study is effective in visualizing RCR damage, but further improvement is required in the evaluation of intermediate damage and the generalization of the evaluation procedure.

Drought is increasingly threatening smallholder farmers in Southeast Asia. The crop insurance system is one of the promising countermeasures that was implemented in Indonesia in 2015. Because the damage assessment in the present system is conducted through direct investigations based on appearance, it is not objective and needs a long time to cover large areas. In this study, we investigated a rapid assessment method for paddy fields using a vegetation index (VI) taken by an unmanned aerial vehicle (UAV) with a multispectral camera in 2019 and 2021. Then, two ways of assessment for drought damage were tested: linear regression (LR) based on a visually assessed drought level (DL), and k-means clustering without an assessed DL. As a result, EVI2 could represent the damage level, showing the tendency of the decrease in the value along with the increasing DL. The estimated DL by both methods mostly coincided with the assessed DL, but the concordance rates varied depending on the locations and the number of assessed fields. Differences in the growth stage and rice cultivars also affected the results. This study revealed the feasibility of the UAV-based rapid and objective assessment method. Further data collection and analysis would be required for implementation in the future.

Understanding the temporal and spatial variability in crop yield is considered as one of the key steps in agricultural risk assessment. Therefore, a study of an irrigated area in Cihea, West Java, Indonesia, was conducted to assess rice yield per field using SENTINEL-2 imagery and yield observation data in 2018 and 2019. The study area is located in the Citarum River basin. SENTINEL-2 images were used to derive paddy rice’s growth curve and estimate rice growth stages based on the normalized difference vegetation index. Using these results, the regression model formula using Band 4 (665 nm) and the normalized difference water index in the ripening stage was created (R2 = 0.40, RMSE = 1.21 t/ha). The results from this model were used to generate yield maps, which illustrated a distinct spatial variation in rice yield, such as the average rice productivity in the study area was relatively high, however, the difference between years tended to be small in the upper stream area. The results of this study show that this method is effective in this area to monitor rice yield condition and distribution.Understanding the temporal and spatial variability in crop yield is considered as one of the key steps in agricultural risk assessment. Therefore, a study of an irrigated area in Cihea, West Java, Indonesia, was conducted to assess rice yield per field using SENTINEL-2 imagery and yield observation data in 2018 and 2019. The study area is located in the Citarum River basin. SENTINEL-2 images were used to derive paddy rice’s growth curve and estimate rice growth stages based on the normalized difference vegetation index. Using these results, the regression model formula using Band 4 (665 nm) and the normalized difference water index in the ripening stage was created (R2 = 0.40, RMSE = 1.21 t/ha). The results from this model were used to generate yield maps, which illustrated a distinct spatial variation in rice yield, such as the average rice productivity in the study area was relatively high, however, the difference between years tended to be small in the upper stream area. The results of this study show that this method is effective in this area to monitor rice yield condition and distribution.

This study aimed to develop a new method for assessing damage to rice plants by pests and diseases using remote sensing data to enable greater efficiency and accuracy for payment of indemnity in the agricultural insurance system of Indonesia, formally operationalized in 2016. The relationships between bacterial leaf blight (BLB) damage ratio in rice crops evaluated by pest observers using the current visual inspection method and the reflectance of each observation band of RapidEye and Sentinel 2, normalized difference vegetation index (NDVI), green NDVI (GNDVI), and red edge multiplied by the green band index (RGI) were studied. The results showed a positive relationship between BLB damage intensity and reflectance of visible wavelength bands, and a particularly strong positive correlation between the red band and BLB damage intensity. The BLB damage intensity can be evaluated based on pixels and paddy parcels. Time series analysis was conducted using Sentinel-2 data acquired during different plant growth periods such as heading, flowering, ripening, maturity, and harvesting. The results showed a strong correlation between the BLB damage intensity and reflectance of the red edge band at the rice heading and flowering stages; the correlation of BLB damage intensity with the reflectance of the visible range became stronger as the rice plant approached the harvesting stage. This study clearly demonstrated that BLB symptoms can be successfully detected and evaluated approximately one or one and a half months before the harvesting period using remote sensing data. We propose that the BLB damage intensity, currently assessed by pest observers through visual inspection methods, can be calculated from satellite data, suggesting that the satellite sensor could play a role similar to the human eye.

本研究ではトウモロコシとダイズをモデル作物として，UAVに搭載された熱赤外カメラを用いて，虫害の早期検出の可能性を検討した．トウモロコシにおいてはアワノメイガの食害が検出可能であったが，早期検出可能であった例は限られ，RGB画像よりも有効であるとは言い切れなかった．ダイズにおいてはハダニにより葉温が明瞭に上昇し，周囲への伝播も観察され，RGB画像より有効であった．以上より，UAV搭載の熱赤外カメラを用いた観測では，吸汁性の虫害検出が効果的であり，他作物への適用も期待できると考えられた．

Bacterial Leaf Blight (BLB) is one of the main diseases in Indonesia that causes a 90% reduction in grain weight. Multispectral imaging may be used as a quick and effective method for damage assessment and is expected to utilize on agricultural insurance in Indonesia. Data were collected at the rainy season and dry season 2018 in Farmers rice field at Bali Province. Vegetation indices (NDVI, GNDVI, and VARI_red-edge) was analyzed using QGIS 2.18 from multispectral images. Some vegetation index shows positive correlation with SPAD and negative correlation with DSI (%). VARI_red-edgehas a higher relationship with DSI (R²: 0.8443) than NDVI (R²: 0.8291) and GNDVI (R²: 0.5463) at the average value on each location, but the relation seems to be affected by that relation between SPAD and LAI. Further data and analysis are required.

In Indonesia, there is a need to improve the efficiency of damage assessments of the agricultural insurance system for paddy rice producers affected by floods, droughts, pests, and diseases. In this study, we develop a method to estimate the transplanting date required for damage assessments of paddy rice fields. The study area is the Cihea irrigation district in West Java, Republic of Indonesia. Backscattering coefficients of VH polarization measured by a synthetic aperture radar onboard the Sentinel-1 satellite were used for the estimations. We investigated the accuracy of the estimations of the proposed method by smoothing out the time-series data, applying a speckle filter, and by signal synthesis of the surrounding fields. It was found that these variations effectively improved the estimation accuracy. To further improve the estimation accuracy, the data for all incident angles were used after correcting the incident angle dependence of the backscattering coefficients for three types of data with different incident angles (32°, 41°, and 45°) obtained in the study area. The estimated transplanting date for each field in the test site was compared with the transplanting date obtained through interviews. The standard deviations of the estimation errors for the four cropping periods from March 2018 to February 2020 were found to be ~5–6 days, and the percentages of estimation errors in transplanting dates within 5, 10, and 15 days were estimated to be 69%, 92%, and 97%, respectively. It was confirmed that a sufficiently reliable transplanting date estimation can be obtained ~10–15 days after transplantation.

It was reported that 326 Japanese Encephalitis (JE) cases in Indonesia in 2016, majority cases (69.3%) occurred in Bali. It shows that Bali is a prone-area to JE incidence. Previous studies noted that JE is closely related to rural and suburban areas where rice culture and pig farming coexist. This study aims at i) determining knowledge and preventive practices of JE by farmer households; (ii) observing types of mosquitos around farmer households; and (iii) mapping the potency of JE spread using geo-spatial information.
 Result from this research shows that farmer households have limited knowledge and preventive actions to the incidence of Japanese encephalitis. Preventive actions carried out by respondents were not for JE incidence as such. Nonetheless, farmer’s response to source of vectors and cleanliness are good preventive actions not just to JE incidence but also for other diseases. In this research, setting up mosquito’s light traps nearby pig pens and rice fields has been successful. The result shows that Culex tritaeniorhynchus was dominant type of mosquito trapped. This is an indication that the selected areas are susceptible to the incidence of JE as the Culex sp was reported as the most competent of JE vector in Asia. Maps of the JE spread in Badung regency also coincidence with the places of Culex sp trapped and the dense of rice field. Based on the limitation of knowledge and preventive actions carried out by farmer households, it is important for the stakeholders in the regency including Regional Health Office and health care workers to socialize the occurrence of JE in the community and how to prevent against the disease.

This study presents a method for detecting rice crop damage due to bacterial leaf blight (BLB) infestation. Rice crop samples are first analyzed using a handheld spectroradiometer. Then, multi-temporal satellite image analysis is used to determine the most suitable vegetation indices for detecting BLB. The results showed that healthy plants have the highest first derivative value of spectral reflectance of the different categories of diseased plants. Significant difference can be found at approximately 690-770 nm (red edge region) which peak or maximum of the first derivative occurs in healthy crop whereas the highest percentage of BLB showed the lowest in that region. Moreover, visible bands such as blue, green, red, and red edge 1 band show variation of correlation in the early (vegetative) to generative stage then getting high especially in early of harvesting stage than the other bands; the NIR band exhibits a low correlation from the early stage of the growing season whereas the red and red edge bands reveal the highest correlations in the later stage of harvesting. Similarly, the satellite image analysis also reveals that disease incidence gradually increases with increasing age of the plant. The vegetation indices whose formulas consist of blue, green, red, and red edge bands (NGRDI, NPCI, and PSRI) exhibit the highest correlation with BLB infestation. NPCI and PSRI indices indicate that crop stress due to BLB is detected from ripening stage of NPCI then the senescence condition is then detected 12 days later. The coefficients of determination between these indices and BLB are 0.44, 0.63, and 0.67, respectively

The trend of rice productivity and its stability in Badung Regency are presented in this study. The area is a tourism-based economy without leaving the role of rice production to feed the people.  Time series data were collected from five sub districts covering 53 villages during 2008-2016 due to the completeness of the data. Variability of the data was observed from the coefficient of variance (C.V.) to decide rice productivity stability. This study also observed correlation between rainfall and rice productivity in the area. Result of the study shows that rice productivity trend in Badung Regency tend to decline during nine years of observation especially in 2014-2016. Three sub-districts had stable condition, while two sub-districts in tourism area contributed to the decline of this matter. Analysis using bi-plot revealed that there is no significant correlation between rainfall in sub-district and rice productivity, implying that water is available throughout the year.  In terms of stability, majority of villages (69.81%) had stable condition of rice productivity ranging from middle and high category. Other 30.19% villages were categorized as unstable ranging from low to high productivity. This condition showed that Badung Regency were able to maintain stability of rice productivity during nine years of observation. Implication of this study is to pay more attention to two sub districts in tourism area to improve their rice productivity such as implementation of good agricultural practices.

Abstract. The objective of this study is to detect flooded area in rice paddy fields in Indonesia by using remotely sensed data. We used synthetic aperture radar (SAR) data for this purpose, because it is capable of getting high-resolution data in almost all-weather conditions. The paper gives a result of detecting flooded area occurred in our research sites located close to Bandung area, Tegalluar in Bojongsoang district, from the end of February to the beginning of March in 2018. The C-band SAR data acquired by Sentinel-1 were used for this analysis. We derived the gamma-naught threshold for dividing flood and non-flood areas by using a linear discriminant analysis. Discriminant accuracy reached 98% using VV polarization. By using the gamma-naught threshold and rice paddy field mask (GIS data), the rice paddy flooded area could be extracted with good accuracy.

Purpose of this study is to estimate rice yield on a plot basis with use of satellite data and field investigation data and to calculate potential quantity of rice straw to be utilized as a raw material for biofuel production in Bali, Indonesia. In addition, for continuous supply of rice straw by farmers to a biofuel producer, it is absolutely necessary to investigate farmer’s interest, behavior and potential issues to solve. For this investigation, an interview was made to the subak heads. The subak is a traditional social organization consisting of farmers for managing irrigation and agriculture in the farmers’ village.The created estimation equation of rice yield had NDVI from SPOT satellite data as a predictor and was significant at 1% level. Based on the rice yield estimated through the equation and the paddy area, quantity of rice straw to be available and quantity of bioethanol to be produced were estimated. In case of Kediri, the rice straw quantity was 42,274 t/year and the ethanol quantity was 12,682 kl.On the other hand, the handling of rice straw after harvesting was leaving/putting back to soil or burning, which accounted for 76% of the rice straw. Concerning a price of rice straw, about 60% of farmers expressed their willingness to sell at 100 rupiah (1 cent US$) per kg, and about 88% including people saying high probability seemed to think this price would be as an adequate level of sales price. Through the interview work, it becomes clear that, when some issues such as price of rice straw are settled, farmers are positive to selling of rice straw for production of biofuel.For realizing bioethanol production from rice straw, there still remain some items to be studied further such as production process, transportation and storage system and costs. The result of our study suggests that a proposal for improvement of stable production can be made through rice yield estimation and monitoring using satellite data and that rice straw can be supplied as a promising resource of raw material for bioethanol production. This is considered to contribute the promotion of activity to reach the national goal of bioethanol production in future.

Human population growth has led to increases in energy and food production, use of fertilizers, and wastewater flows. Enhanced availability of nitrogen is a cause of eutrophication of rivers, lakes, and estuaries worldwide. In this study, Citarum River Basin, West Java, Indonesia was selected as a target area, supplying 80% of domestic water to Jakarta Metropolis. Meteorological and hydrological data from 1996 to 2009, and spatial data such as topography, land use, soil properties were collected for model simulation. Conceptual nitrogen balance model, which has three nitrogen pools, was developed and combined with rainfall runoff model. Proposed model was applied to the Citarum River Basin and simulated river discharge and nitrogen load in 1 km X 1 km resolution to check the model applicability. By using the model, spatial distribution of nitrogen loading in whole basin level was estimated and histograms of nitrogen load from different land use also evaluated. The results provide a first insight into the magnitude and the spatial distribution of nitrogen loading in Citarum River Basin.

<p>Goal is to develop new method utilizing satellite data for assessment of damage in paddy field which can contribute toward substantial reduction of the damage assessment time and costs in framework of agricultural insurance. For the damage assessment, estimation of yield in each paddy plot is a key, so the research on the estimation of rice yeild was carried out using satellite data in Hokkaido, Japan. Both multiple linear regression analysis and the projection pursuit regression analysis were conducted for the estimation of yeild using data from 3 different satellites about 3 different rice varieties. As the result, the projection pursuit regression analysis showed smaller value of predictive error than that of the multiple linear regression analysis, and the lowest predictive error was indicated when SPOT5 data with 10 m resolution was used. Moreover, for reducing the predictive error, it was found that a yield estimation formula should be created for each of different rice varieties. The results out of this research suggest that satellite data can be effectively used for estimation of yield and also assessment of damage at lower costs to calculate indemnity in agricultural insurance.</p>

北海道十勝地域の黒ボク土地帯を対象とし,未耕地と耕地のA層内の土壌炭素量を比較し,耕地化による土壌炭素量の増減を検討した.(1)耕地化による土壌炭素増減から,耕地化により土壌炭素が増加したグループ1(淡色黒ボク土),変化の少ないグループ2(多腐植質普通黒ボク土),減少したグループ3(多腐植質普通黒ボク土,多腐植質普通多湿黒ボク土,多腐植質厚層多湿黒ボク土)に大別された.グループ1では耕地化により21.9-31.8MgCha^<-1>の土壌炭素量が増加していたのに対して,グループ3では耕地化により78.5-149.8MgCha^<-1>の土壌炭素量が減少していた.(2)グループ3では未耕地の土壌炭素量が250MgCha^<-1>以上と他のグループに比べて高く,暗渠排水が積極的に施工されている特徴が見られた.

The impact of climate change on food production in Indonesia is anticipated to be a level where food production in 2050 will decline ranging from 38% to more than ten-folds of the current production, and it is necessary to improve a cultivation management system aiming at increasing rice yield and also the stability. For this improvement, effective use of irrigation water and water distribution management plan is critical and has to be carefully reviewed and designed. In this study, the amount of irrigation water necessary for rice cultivation at each stage of its growth was calculated and reviewed using satellite data acquired in dry cultivation season for the paddy field of West Java, Indonesia. The results of the study make it clear that the new method developed in the study can calculate and estimate the water amount to be required at each growth stage of rice; the requirement amount can be visualized by each pixel of satellite data; also, the change of the water requirement amount can be checked on almost real-time basis.

Existing methods for rice field classification have some limitations due to the large variety of land covers attributed to rice fields. This study used temporal variance analysis of daily Moderate Resolution Imaging Spectroradiometer (MODIS) satellite images to discriminate rice fields from other land uses. The classification result was then compared with the reference data. Regression analysis showed that regency and district comparisons produced coefficients of determination (R-2) of 0.97490 and 0.92298, whereas the root mean square errors (RMSEs) were 1570.70 and 551.36 ha, respectively. The overall accuracy of the method in this study was 87.91%, with commission and omission errors of 35.45% and 17.68%, respectively. Kappa analysis showed strong agreement between the results of the analysis of the MODIS data using the method developed in this study and the reference data, with a kappa coefficient value of 0.8371. The results of this study indicated that the algorithm for variance analysis of multitemporal MODIS images could potentially be applied for rice field mapping.

The standard image classification method typically uses multispectral imageryon one acquisition date as an input for classification. Rice fields exhibit high variability inland cover states, which influences their reflectance. Using the existing standard method forrice field classification may increase errors of commission and omission, thereby reducingclassification accuracy. This study utilised temporal variance in a vegetation index as amodified input image for rice field classification. The results showed that classification ofrice fields using modified input images provided a better result. Using the modifiedclassification input improved the correspondence between rice field area obtained from theclassification result and reference data (R2 increased from 0.2557 to 0.9656 for regencylevelcomparisons and from 0.5045 to 0.8698 for district-level comparisons). Theclassification accuracy and the estimated Kappa value also increased when using themodified classification input compared to the standard method, from 66.33 to 83.73 andfrom 0.49 to 0.77, respectively. The commission error, omission error, and Kappa variancedecreased from 68.11 to 42.36, 28.48 to 27.97, and 0.00159 to 0.00039, respectively, whenusing modified input images compared to the standard method. The Kappa analysisconcluded that there are significant differences between the procedure developed in thisstudy and the standard method for rice field classification. Consequently, the modifiedclassification method developed here is significant improvement over the standardprocedure.

Soil carbon stocks (SCSs) in an upland area of Tokachi District, Hokkaido, Japan, were estimated by satellite remote sensing and a soil survey. The soil parent materials in the studied area were alluvial deposits, volcanic ash plus alluvial deposits, and volcanic ash. Surface soil carbon concentrations (SSCCs) were negatively correlated with satellite image data (green, red, and near-infrared reflectance) for each parent material. The highest correlations between reflectance and SSCCs were obtained from red wavelength reflectance for alluvial deposits (r=-0.82, p<0.01) and volcanic ash plus alluvial deposits (r=-0.91, p<0.01), and from near-infrared reflectance for volcanic ash (r=-0.90, p<0.01). We generated an SSCC map of the study area using the regression equations and satellite reflectance data. The soil survey results showed that SSCCs were positively correlated with SCSs in the 0-30cm depth interval for each parent material (best-fit regressions: alluvial deposits, r=0.97; volcanic ash plus alluvial deposits, r=0.97; volcanic ash, r=0.97), and they were also positively correlated with SCSs in the 30-95cm depth interval for volcanic ash (r=0.94). We were therefore able to generate a map of estimated SCSs from the SSCC map and the regression equations developed between SSCCs and SCSs. The estimated and measured SCSs in both depth intervals showed an almost 1:1 relationship, with root-mean-square errors of 19.5 (0-30cm) and 28.6Mg carbon (C) ha^<-1> (30-95cm). According to the SCS maps, SCSs at 0-30cm depth in areas of alluvial deposits, volcanic ash plus alluvial deposits, and volcanic ash were mostly 50-150, 100-200, and 50-250MgCha^<-1>, respectively, whereas at 30-95cm depth in the volcanic ash area they were mostly less than 250MgCha^<-1>.