Rice Paddy Area Research Articles

Correlation Analysis of Vegetation Index Impact on Rice Paddy Productivity Estimation using Landsat-8 and Sentinel-2A Images (Case Study: Blitar District)

Abstract Rice is consumed by most of the population in Indonesia. The generative phase, the final stage of rice growth cycle, produces raw grains that are processed into rice, forming the basis of rice production. To estimate rice paddy productivity in March 2023, vegetation indices such as the Normalized Difference Vegetation Index (NDVI) and the Optimized Soil-Adjusted Vegetation Index (OSAVI) are used. These indices are crucial as they correlate with rice paddy areas in the generative phase. Sentinel-2A imagery calculations have shown strong correlations, with NDVI at R = 0.61 and OSAVI at R = 0.75. The OSAVI index from Sentinel-2A yielded an average productivity estimate of 31.35 ton/ha with an RMSE of 5.7041, while NDVI gave an average of 63.63 ton/ha with an RMSE of 12.3224. In comparison, Landsat-8 imagery estimates using NDVI showed a moderate correlation (R = 0.54), with an average productivity of 20.75 ton/ha and an RMSE of 3.2321. The OSAVI index with Landsat-8 provided an average estimate of 64.27 ton/ha with an RMSE of 12.3338 and a moderate correlation (R = 0.41). The correlation coefficients from Sentinel-2A are higher than those from Landsat-8, indicating that the choice of imagery significantly affects rice productivity estimates.

Prediction and Simulation for Land Use and Land Cover Change of Paddy Field Influence by Salinization in Coastal Demak Regency

The extent of coastal rice paddy agricultural land is vulnerable to land use and land cover (LULC) changes to non-agricultural uses due to land degradation, one of which is caused by salinity. This study aims to detect and project LULC changes up to 2031, particularly in coastal rice paddy areas affected by salinity, by comparing LULC in 2017, 2019, and 2021. Sentinel-2 Imagery is used for LULC classification, with recordings selected during the generative phase of rice growth to obtain the most optimal rice paddy area. There are six LULC classifications: water, wetland, low-medium-high vegetation cover, and built-up area. To understand the impact of salinity on crops, several vegetation indices (VIs) such as NDVI, SAVI, EVI, and ARVI are used. The LULC changes classified according to VIs are compared with the MOLUSCE plugin based on artificial neural networkmultilayer perceptron (ANN-MLP) and Cellular Automata (CA). The comparison of VIs results shows that NDVI is better at describing LULC changes due to the influence of salinity, with a kappa value of 0.63 and a Correctness of 72.565. The LULC projection using CA in all VIs indicates that wetland areas are more likely to convert into water bodies, suggesting that high salinity land tends to be unproductive for rice paddies, making it prone to conversion.

Assessing paddy methane emissions through the identification of rice and winter crop areas using Sentinel-2 imagery in Korea

The global efforts on reducing methane (CH4) emissions was emphasized in COP 28 and the potential for improved estimation became feasible through bottom-up data acquisition with advanced remote sensing technology. The objectives of this study were to extract summer rice and winter crop cultivation areas based on satellite images and to incorporate into estimating CH4 emissions in South Korea for the year 2020. Satellite images of Sentinel-1 and Sentinel-2 were acquired from European Space Agency. Rice paddy was classified with backscattering coefficient from Sentinel-1 images, while the normalized difference vegetation index from Sentinel-2 images was used to identify winter cropping field. The equation of IPCC guidelines was used to estimate CH4 emissions by incorporating the areas of rice paddy and winter crop extracted with the respective satellite image. National farming statistics were used to determine the scaling factors for paddy organic matter and water management practices. The estimated areas for rice paddy and winter crop cultivation were 712,237 ha and 117,840 ha, respectively. The rice paddy areas were primarily concentrated in the western regions of the Korean peninsula, whereas winter crop cultivation was predominantly found in southern part of the country. The total amount of CH4 emissions was 6272 Gg CO2 eq./yr when considering rice straw and winter cropping practices into estimation (modified Tier 2 method). This represents a 7% increase compared to the method that considered solely the rice straw incorporation (current Tier 2 method). The CH4 emissions per unit area were also 8.82 tons CO2 eq./ha/yr with the modified Tier 2 method, indicating a 10% greater compared to the current Tier 2 method. Substantial CH4 emissions were primarily concentrated in western regions where extensive rice paddy cultivation occurs, while greater CH4 emissions per unit area were predominantly found in southern regions with substantial winter crop cultivation. The study findings hold importance for improving the accuracy of CH4 emissions estimation by employing bottom-up approach that utilizes satellite imagery to assess rice paddy and winter cropping areas. Further study would be needed to incorporate field-based data on rice crop management practices, such as rice straw and water management, to further refine CH4 emission estimation method.

Behavioral responses to predator and heterospecific alarm calls are habitat-specific in Eurasian tree sparrows

Acoustic communication plays a vital role in predator-prey interactions. Although habitat structure has been shown to affect anti-predator tactics, little is known about how animals vary their behaviors in response to predator calls or heterospecific alarm calls in different environments. Here we used sound playbacks to test the responses of Eurasian tree sparrows (Passer montanus) foraging in harvested/unharvested rice paddy and open residential area. In the first experiment, we tested their behavioral responses to dove calls, male common cuckoo (Cuculus canorus) calls, hawk-like calls mimicked by female common cuckoo, sparrowhawk (Accipiter nisus) calls, and human yell calls produced to scare birds (predator signal playbacks). In the second experiment, we tested their behavioral responses to the Japanese tit’s (Parus minor) territorial songs and alarm calls (heterospecific alarm signal playbacks). Results showed that the tree sparrows had less fleeing in unharvested ripe rice paddy than in harvested rice paddy and open residential area. In predator signal playbacks, call type affected the escape behavior of sparrows in unharvested rice paddy and open residential area but not harvested rice paddy. In alarm signal playbacks, tit alarm calls evoked more fleeing than territorial songs in harvested rice paddy and open residential area but not unharvested rice paddy. These results suggest that anthropogenic habitat changes may influence avian anti-predator tactics.

Comparison of Cloud-Mask Algorithms and Machine-Learning Methods Using Sentinel-2 Imagery for Mapping Paddy Rice in Jianghan Plain

Southern China, one of the traditional rice production bases, has experienced significant declines in the area of rice paddy since the beginning of this century. Monitoring the rice cropping area is becoming an urgent need for food security policy decisions. One of the main challenges for mapping rice in this area is the quantity of cloud-free observations that are vulnerable to frequent cloud cover. Another relevant issue that needs to be addressed is determining how to select the appropriate classifier for mapping paddy rice based on the cloud-masked observations. Therefore, this study was organized to quickly find a strategy for rice mapping by evaluating cloud-mask algorithms and machine-learning methods for Sentinel-2 imagery. Specifically, we compared four GEE-embedded cloud-mask algorithms (QA60, S2cloudless, CloudScore, and CDI (Cloud Displacement Index)) and analyzed the appropriateness of widely accepted machine-learning classifiers (random forest, support vector machine, classification and regression tree, gradient tree boost) for cloud-masked imagery. The S2cloudless algorithm had a clear edge over the other three algorithms based on its overall accuracy in evaluation and visual inspection. The findings showed that the algorithm with a combination of S2cloudless and random forest showed the best performance when comparing mapping results with field survey data, referenced rice maps, and statistical yearbooks. In general, the research highlighted the potential of using Sentinel-2 imagery to map paddy rice with multiple combinations of cloud-mask algorithms and machine-learning methods in a cloud-prone area, which has the potential to broaden our rice mapping strategies.

Land use change from paddy rice to sugarcane under long-term no-till conditions: increase P balance, soil organic matter and sugarcane productivity

ABSTRACT Expanding sugarcane (SC) cultivation at the expense of paddy rice (PRC) areas led to soil nutrient cycling changes. The study hypothesized that SC straw return and fertilizer application after land use change (LUC) from PRC to SC cropping would increase the soil-crop systems nutrient balance, ultimately enhancing growth, yield, and nutrient use efficiency of SC. In a field study involving a 17-year chronosequence of sugarcane 6 (SC6), 13 (SC13), and 17 (SC17) years after LUC, the nutrient balance dynamics of SC were monitored in comparison with those of PRC. A phosphorus (P) balance increase of 62–102% after LUC increased soil available P (Pav) from 5.2 to 13.6–16.0 mg kg−1. A potassium (K) balance decrease of 47–55% after LUC led to a decrease in soil exchangeable K (Kex) from 102.5 to 96.2–56.8 mg kg−1. Moreover, LUC increased the % sand, soil organic matter (SOM), total mineral nitrogen (N), and total N. In contrast, the % clay, bulk density, cation exchange capacity (CECpH7), and exchangeable calcium (Caex) decreased after LUC. An increase in soil nutrients increased cane yield and nutrient use efficiency with time after LUC. However, further research in different locations and management practices is required for the LUC investigation.

Evaluation of Korean methane emission sources with satellite retrievals by spatial correlation analysis

Methane is a significant greenhouse gas (GHG), and it is imperative to understand its spatiotemporal distribution and primary sources in areas with higher methane concentrations, as such insights are essential for informing effective mitigation policies. In this study, we employed TROPOMI satellite retrievals to analyze the spatiotemporal patterns of methane distributions and identify major emission sources in South Korea over the period from August 2018 to July 2019. Additionally, we examined the spatial correlations between satellite methane retrievals and emission sources to characterize regions with higher methane levels on an annual basis.Concerning spatial distributions, concentrations exceeding 1870 ppb were predominantly observed in western non-mountainous regions, particularly in rice paddy areas. Moreover, sporadic concentrations exceeding 1880 ppb were detected in large ports and industrial zones, primarily located in coastal regions of South Korea.Our spatial correlation analysis, conducted using the SDMSelect method, identified specific emissions contributing to regions with higher methane concentrations. There were some areas with relatively strong correlations between high XCH4 and emissions from the domestic livestock industry, fossil fuel utilization (specifically, the oil and gas sector), landfills, and rice paddies. This analysis, incorporating domestic emission inventories and satellite data, provides valuable insights into the characteristics of regional methane concentrations. In addition, this analysis can assess national methane emissions inventories, where there is limited information on the spatial distributions, offering critical information for the prioritization of domestic regional policies aimed at reducing greenhouse gas emissions.

Evaluating Landsat-9 TIRS-2 calibrations and land surface temperature retrievals against ground measurements using multi-instrument spatial and temporal sampling along transects

High-resolution TIR data are essential for a wide range of studies, e.g., in environmental monitoring, urban effects, water stress, agricultural productivity, and land/water resources. Landsat series provides high-resolution TIR data, but data accuracy must be ensured to contribute to these fields. After the calibration problems of Landsat 8 (L8) TIRS due to a stray light effect, efforts were made in the design of Landsat 9 (L9) TIRS-2 to mitigate this effect. In any case, field calibration and validation activities are necessary to evaluate the accuracy of recently launched sensors. These activities require high-quality ground TIR radiance data that are representative at satellite spatial resolution and that quantify spatial–temporal fluctuations at the experimental sites concurrently to satellite overpasses. However, these requirements are not always accomplished in the literature.This paper provides vicarious calibrations of the L9 TIRS-2 data with accurate ground measurements acquired using multi-instrument spatial and temporal sampling in a rice paddy area from December 2021 to August 2023, in order to contribute to the calibration and validation of L9 TIRS-data.In March 1, 2023 L9 TIRS-2 scenes started to be reprocessed and calibration was updated. Both the original calibration and that after reprocessing were evaluated. Close results were obtained for band 10 (b10). When comparing satellite brightness temperatures and those determined from in situ LSTs, systematic differences of −0.3 K were shown, with root-mean-square differences (RMSDs) of 0.5 K. Results for band 11 showed a bias of −0.1 K for the original data and of 0.2 K after the reprocessing, with RMSDs of 0.6 K.The L2 LST product generated operationally from b10 data was also evaluated, together with alternative single-channel (SC) algorithms. A bias (RMSD) of 0.4 K (1.1 K) was obtained for the product. However, negative biases were shown both with the radiative transfer equation used on b10 radiances and the alternative SC algorithm.Finally, we validated six types of split-window (SW) algorithms applied to L9 TIRS-2 data before and after the reprocessing. A systematic difference of about −0.6 K between SW LST biases when using the reprocessed and original data was observed. In any case, for most of the algorithms with original and reprocessed data, LST biases and standard deviations were within the recommended threshold of 1 K, which proves the good performance of L9 TIRS-2 data to monitor LST at high spatial resolution.

Comparison of Different Supervised Classification Algorithms for Mapping Paddy Rice Areas Using Landsat 9 Imageries

Rice is known to be one of the most essential crops in Turkey, as well as many other countries especially in Asia, whereas paddy rice cropping systems have a key role in many processes ranging from human nutrition to environment-related perspectives. Therefore, determination of cultivation area is still a hot topic among researchers from various disciplines, planners, and decision makers. In present study, it was aimed to evaluate performances of three classifications algorithms among most widely used ones, namely, maximum likelihood (ML), random forest (RF), and k-nearest neighborhood (KNN), for paddy rice mapping in a mixed cultivation area located in Biga District of Çanakkale Province, Turkey. Visual, near-infrared and shortwave infrared bands of Landsat 9 acquired in dry season of 2022 year was utilized. The classification scheme included six classes as dense vegetation (D), sparse vegetation (S), agricultural field (A), water surface (W), residential area – base soil (RB), and paddy rice (PR). The performances were tested using the same training samples and accuracy control points. The reliability of each classification was evaluated through accuracy assessments considering 150 equalized randomized control points. Accordingly, RF algorithym could identify PR areas with over 96.0% accuracy, and it was followed by KNN with 92.0%.

Inference of selection pressures that drive insecticide resistance in Anopheles and Culex mosquitoes in Korea

Pyrethroids are primarily used for mosquito control in Korea. However, high frequencies of mutations conferring resistance to not only pyrethroids but also to other insecticides have been found in mosquito populations. This study aimed to examine the hypothesis that insecticides used outside of public health may play a role in selection. Briefly, the resistance mutation frequencies to three insecticide groups (pyrethroids, organophosphates, and cyclodienes) were estimated in two representative groups of mosquito species (Anopheles Hyrcanus Group and Culex pipiens complex). The relationship between these frequencies and the land-use status of the collection sites was investigated through multiple regression analysis. In the Anopheles Hyrcanus Group, the frequencies of both ace1 (organophosphate resistance) and rdl (cyclodiene resistance) mutations were positively correlated with ‘proximity to golf course’, possibly be due to the insecticides used for turf maintenance. They also showed positive correlations with field area and rice paddy area, respectively, suggesting the role of agricultural insecticides in the selection of these resistance traits. For the Cx. pipiens complex, the kdr (pyrethroid resistance), ace1, and rdl mutations were positively correlated with the residential area, field, and rice paddy, respectively. Therefore, pyrethroids used for public health could serve as a direct source of resistance selection pressure against kdr, whereas non-public health insecticides may pose primary selection pressure against the ace1 and rdl traits. The current findings suggest that the insecticides used in agriculture and the golf industry play a significant role in mosquito selection, despite variations in the extent of indirect selection pressure according to the mosquito groups and insecticide classes.

Paddy Rice Phenological Mapping throughout 30-Years Satellite Images in the Honghe Hani Rice Terraces

The Honghe Hani Rice Terraces represent the coexistence between natural and cultural systems. Despite being listed as a World Heritage Site in 2013, certain natural and anthropogenic factors have changed land use/land cover, which has led to a reduction in the size of the paddy rice area. It is difficult to accurately assess these changes due to the lack of historical maps of paddy rice croplands with fine spatial resolution. Therefore, we integrated a random forest classifier and phenological information to improve mapping accuracy and stability. We then mapped the historical distribution of land use/land cover in the Honghe Hani Rice Terraces from 1989–1991 to 2019–2021 using the Google Earth Engine. Finally, we analyzed the driving forces of land use types in the Honghe Hani Rice Terraces. We found that: (1) forests, shrubs or grasslands, and other croplands could be discriminated from paddy rice during the flooding and transplanting period, and water bodies and buildings could also be discriminated from paddy rice during the growing and harvesting period. (2) Inputting phenological feature data improved mapping accuracy and stability compared with single phenological periods. (3) In the past thirty years, 10.651%, 8.810%, and 5.711% of paddy rice were respectively converted to forests, shrubs or grasslands, and other croplands in the Honghe Hani Rice Terraces. (4) Lower agricultural profits and drought led to problems in identifying the driving mechanisms behind paddy rice distribution changes. This study demonstrates that phenological information can improve the mapping accuracy of rice terraces. It also provides evidence for the change in the size of the rice terrace area and associated driving forces in Southwest China.

Twenty-meter annual paddy rice area map for mainland Southeast Asia using Sentinel-1 synthetic-aperture-radar data

Abstract. Over 90 % of the world's rice is produced in the Asia–Pacific region. Synthetic-aperture radar (SAR) enables all-day and all-weather observations of rice distribution in tropical and subtropical regions. The complexity of rice cultivation patterns in tropical and subtropical regions makes it difficult to construct a representative data-relevant rice crop model, increasing the difficulty in extracting rice distributions from SAR data. To address this problem, a rice area mapping method for large regional tropical or subtropical areas based on time-series Sentinel-1 SAR data is proposed in this study. Based on the analysis of rice backscattering characteristics in mainland Southeast Asia, the combination of spatiotemporal statistical features with good generalization ability was selected and then input into the U-Net semantic segmentation model, combined with WorldCover data to reduce false alarms, finally the 20 m resolution rice area map of five countries in mainland Southeast Asia in 2019 was obtained. The proposed method achieved an accuracy of 92.20 % on the validation sample set, and the good agreement was obtained when comparing our rice area map with statistical data and other rice area maps at the national and provincial levels. The maximum coefficient of determination R2 was 0.93 at the national level and 0.97 at the provincial level. These results demonstrate the advantages of the proposed method in rice area mapping with complex cropping patterns and the reliability of the generated rice area maps. The 20 m annual paddy rice area map for mainland Southeast Asia is available at https://doi.org/10.5281/zenodo.7315076 (Sun et al., 2022b).

Transformative change of paddy rice systems for biodiversity: A case study of the crested ibis certified rice system in Sado Island, Japan

ABSTRACT Innovation history analysis of the transformation of paddy rice system on Sado Island from 1999 to 2019, which progressed with the reintroduction of crested ibis (Nipponia nippon), identified three key levers, i.e. multi-stakeholder collaboration, integrated local policies, and incentives, which, interacting one another and with seven leverage points and six exogenous factors, drove the development and deployment of the Crested Ibis Rice Certification scheme, resulting in a quarter of all staple rice paddy area on the island being certified by 2016. The results suggest two potential levers to further expand the certified area beyond this equilibrium for a wider transformation.

Remote sensing monitoring of maize and paddy rice planting area using GF-6 WFV red edge features

In order to obtain the high-resolution crop planting information of Liaoning Province in Northeast China and to explore the crop extraction ability of the red edge features of GF-6 satellite at the provincial scale. Five red edge features of RE1, RE2, NDRE, NDVIre1 and NDVI710 were extracted from four periods of GF-6 WFV images from May to September 2020, according to the temporal changes of the ground objects, multiple decision tree models were constructed and the random forest method was used to extract the planting area of maize and paddy rice in the study area. The results showed that the woodland and grassland could be extracted with NDRE, NDVIre1 and NDVI710 in mid-early June, while the water and construction land could be extracted with NDRE, NDVIre1 and RE2 from late July to late September. Maize is more conducive to remote sensing monitoring in the late stage of the growing season, such as heading stage and maturing stage, while paddy rice is more easily identified by image in the late transplanting stage and early reviving stage. In terms of spatial distribution, maize and paddy rice are mainly planted in areas with relatively flat terrain, gentle slope, high temperature and abundant precipitation, the maize is widely distributed in plain, mountainous and hilly areas, and paddy rice is mainly planted in the Liaohe Plain. The overall accuracy and kappa coefficient based on the verification sample points were 97.6% and 0.952, respectively. The user accuracy and producer accuracy of maize were 99.5% and 97.5%, and the user accuracy and producer accuracy of paddy rice were 100% and 97.8%, respectively. The extraction area of maize and paddy rice in Liaoning Province were 2787.39 × 103 hm2 and 487.75 × 103 hm2, respectively. Compared with the statistical yearbook data, the extraction accuracy of maize and paddy rice reached 96.7% and 93.7% respectively. The results indicated that only based on GF-6 WFV temporal red edge features, maize and paddy rice in Liaoning Province can be accurately identified. This method is technically feasible and can achieve high resolution and high precision classification results on a large provincial scale.

Agricultural crop pattern mapping and change analysis at a sub-district level in South-eastern region, Bangladesh using Landsat satellite data from 2010 to 2019

The study first time identified and analyzed winter season agricultural crop patterns (ACP) derived from Land use (LU) maps in between 2010 to 2019 of south-eastern regions of Chittagong, Bangladesh. ACP identification was a challenging task in the worldwide research relevant to crop-related studies. To overcome this, we have considered frequently used traditional unsupervised classifier, such as K-means clustering algorithm technique. This has been applied on 30m pixel Landsat satellite reflectance images to identify crop pattern of the study area using the ENVI 5.3 and ArcGIS 10.8 software’s, respectively. Multiple crops with seven classes were identified with the validation of in-situ ground-truth data and Google Earth (GE) images. The overall accuracy and kappa coefficient values were found at 81.96% and 0.79, respectively. The results suggest a significant variation of crop patterns in the study area and in recent time, the area largely dependent on mixed irrigation approach. Moreover, the crop pattern change was observed in the studied period as mixed crop 19% (9282.17 ha), Lentils (Pelon) 24.80% (11594.38 ha), Melon (Bangi) 22.37% (10461.08 ha), Chilis 17.90% (8367.48 ha), Paddy rice, unused land, and other crops, respectively. Among them, Lentils (Pelon) and Melon (Bangi) are identified as two common crops followed by mixed crops category, cultivated in the winter season as it required less irrigation compared to paddy rice area

Global food crop redistribution reduces water footprint without compromising species diversity

Food crop yield and species diversity are critical for global population growth, undernourishment, and environmental burden. Crop redistribution is a practical solution to the sustainable intensification of agricultural systems by replacing existing crops with more suitable ones. Here we test this redistribution theory and its impacts on species diversity using six decades of data on the annual world production of 137 food crop products and their average water footprint (WFP). We find the global food crop effective species diversity had an increasing trend from 21 to 27 in recent six decades. Then, we develop a nutrient water footprint productivity index (WFPPI) to optimize crop distribution. Large calorie (+14%), mineral (+25%), and vitamin (+26%) production increases are possible from reshaping the global distribution of 100 crops within the current total WFP and area harvested. Wheat and paddy rice area can reduce by 31% and 34%, respectively, while increasing maize and soybean area by 18% and 40%, respectively. The global redistribution could feed an additional 866 million people while supplying 37% more than the caloric demand for the projected biofuel growth between 2020 and 2030 in the Net Zero Scenario. Such an optimization process does not entail a loss of crop species diversity. With a change of no more than 5%, the stability of the total area harvested for the four largest crops (wheat, maize, paddy rice, and soybeans) is critical for both production and species diversity of all crops. Our results suggest that global crop redistribution provides a promising solution to sustainable production.

Understanding Recovery Is as Important as Understanding Decline: The Case of the Crested Ibis in China

The wild population of the crested ibis (Nipponia nippon) has recovered remarkably from seven individuals in 1981 to over 7000 in 2021. However, it is unclear how key factors, from endogenous density dependence to exogenous environmental pressure, have contributed to the species’ recovery. We used species distribution models to quantify the contributions of climatic variables, human impact, land form and land use in order to understand the recovery process in the context of prevailing environmental conditions. We also calculated the nest density over the past 39 years to estimate the influence of density dependence on population dynamics. We found that the interaction between rice paddy areas and water bodies (rivers, lakes and ponds) had the highest contribution to nest site selection, whereas linear terms for either rice paddies or water bodies alone had little effect. During its recovery, sub-populations in two watersheds have been constrained by high density and have experienced logistic growth, while other sub-populations in over seven watersheds are growing exponentially. Our models indicate that exogenous environmental factors are more important than density restriction at this stage. In China’s transformed landscape, the crested ibis needs both rice paddies and water bodies to fulfil its annual life cycle. Habitat protection should thus cover both habitat types to ensure the long-term survival of this still endangered species.

Hydrological control of the chemical characteristics of suspended particulate phosphorus in the Yasu River watershed, Japan: Implications for its source and bioavailability

AbstractSeveral studies have demonstrated the importance of the bioavailability of particulate phosphorus (PP) exported from various watersheds, especially agricultural land. However, few studies have documented the influence of hydrological conditions on the sources and bioavailability of PP exported from watersheds with multiple land uses. In this study, to assess the sources and bioavailability of PP under different hydrological conditions, we conducted synoptic monitoring of compositional changes in the chemical fractions of PP exported from the Yasu River Watershed, which is covered with a mixture of forest (57.7%) and rice paddy (20.5%) areas. In this river, total phosphorus was composed mainly of PP at all monitoring sites and time points. The sequential extraction of PP revealed that a fraction bound to redox‐sensitive iron and manganese, regarded as bioavailable, was specific to the drainage waters of rice paddies, while a fraction bound to organic matter, regarded as less bioavailable, was specific to a forested stream. In the downstream of the Yasu River covering the whole watershed, PP exported under the base flow and small rainfall conditions was composed mainly of the rice paddy‐derived fraction, while most of the PP exported under heavy rainfall conditions was the forest‐derived fraction. These results suggest that the sources and bioavailability of PP can vary dynamically, depending on the hydrological regimes and land use pattern in the watershed. Even taking into account the fact that the large amount of PP exported during heavy rainfall is considered to be forest‐derived PP with low bioavailability, the monitoring results suggested that the annual amount of bioavailable PP exported is larger than that of dissolved phosphorus, and can have a large impact on phosphorus recycling in the downstream lake ecosystem through sedimentation and regeneration.

Evaluating the necessity of geographical locality for patterning biological integrity and its responses to multiple stressors in river systems

Water and sediment quality influence biotic communities, and relationships between the chemical composition and community structure provide important information about the regulation of river systems. We evaluated the effectiveness of clustering based on stress-response relationships and relationships with geographical location for the appropriate classification of rivers and streams. To address this issues, we applied clustering results using a self-organizing map (SOM) and geo-self-organizing map (Geo-SOM) as random effects in linear mixed-effect models (LMMs). The dataset was composed of 16 water quality variables and 11 sediment quality variables with indices of benthic diatoms (TDI), macroinvertebrates (BMI), and fishes (FAI) surveyed at 84 stations along rivers of the Republic of Korea over 2 years. The clustering results of SOM were spatially sporadic based on the relationships between the variables, while those of Geo-SOM were based on these relationships as well as the geographical characteristics. The inclusion of random effects of SOM and Geo-SOM clustering improved the performance of all LMMs; however, the LMMs with random effects of Geo-SOM-based clustering were best for all three indices. Based on the variable importance from the Geo-SOM training and LMMs, the most influential water quality variables were phosphate and chlorophyll a, and the most influential sediment quality variables were sedimentary phosphate and Cr. The correlation coefficients for heavy metals from the Geo-SOM training and contribution to selected LMMs were generally lower than those of nutrient-associated variables, indicating that the rivers and streams of Korea were more influenced by organic pollution than by heavy metal pollution. Cluster-specific conditions were significantly affected by land-use patterns, showing negative relationships with the percentage of rice paddy area and positive relationships with the percentage of forest. This approach can be applied to the ecological classification of river basins and provides fundamental information for the implementation of sustainable management.

Mapping Paddy Rice in Complex Landscapes with Landsat Time Series Data and Superpixel-Based Deep Learning Method

The spatial pattern and temporal variation in paddy rice areas captured by remote sensing imagery provide an effective way of performing crop management and developing suitable agricultural policies. However, fragmented and scattered rice paddies due to undulating and varied topography, and the availability and quality of remote sensing images (e.g., frequent cloud coverage) pose significant challenges to accurate long-term rice mapping, especially for traditional pixel and phenological methods in subtropical monsoon regions. This study proposed a superpixel and deep-learning-based time series method to analyze Landsat time series data for paddy rice classification in complex landscape regions. First, a superpixel segmentation map was generated using a dynamic-time-warping-based simple non-iterative clustering algorithm with preprocessed spectral indices (SIs) time series data. Second, the SI images were overlaid onto the superpixel map to construct mean SIs time series for each superpixel. Third, a multivariate long short-term memory full convolution neural network (MLSTM-FCN) classifier was employed to learn time series features of rice paddies to produce accurate paddy rice maps. The method was evaluated using Landsat imagery from 2000 to 2020 in Cengong County, Guizhou Province, China. Results indicate that the superpixel MLSTM-FCN achieved a high performance with an overall accuracy varying from 0.9547 to 0.9721, which presents an 0.17–1.23% improvement compared to the random forest method. This study showed that combining spectral, spatial, and temporal features with deep learning methods can generate accurate paddy rice maps in complex landscape regions.