Methane (CH₄) produced in rice paddy soil is transported to the atmosphere mainly via the rice plants and partly by bubbling events (ebullition). Recent studies have shown that ebullition is more significant than previously thought in fields planted with the popular Japanese cultivar ‘Koshihikari’. It remains unclear whether the substantial contribution of ebullition is unique to this specific cultivar, as no previous reports have compared plant-derived and bubbling fluxes separately among various cultivars. Therefore, we planted 22 genetically diverse rice cultivars and measured plant-mediated and bubbling fluxes at three growth stages. Both fluxes, as well as the contribution of bubbling to the total flux, differed among the cultivars. The plant-mediated flux in Koshihikari was similar to or less than those in other cultivars, whereas the bubbling flux and its contribution to total flux were larger, especially at the later stage. The absence of a correlation between plant-mediated flux and dissolved CH₄ in the soil water at the later stage suggests that varietal differences in CH₄ entry from the soil to the plant or gas flow permeability in the plant, rather than the pool size of CH₄ in the soil, control the plant-mediated flux. On the other hand, the increase in bubbling flux associated with plant maturation and its close correspondence with dissolved CH₄ concentration indicate that bubbling flux was controlled by the size of CH₄ pool in the soil, which likely increased with senescence and decay of rice roots. A low correspondence between panicle weight and CH₄ emissions points to the potential for breeding high-yielding rice cultivars with low CH₄ emissions

　Evaporative loss is a substantial fraction of the water budget of lakes. Quantifying lake evaporation is crucial for water resources management, especially in the face of a changing climate and increasing water demand. The recently developed Surface-Temperature and Available-Energy-Based Lake Evaporation (STAEBLE) model implements a new mass and heat transfer method that does not require locally measured evaporation for calibration and which has potential for deployment in any geographical location. In its original article, STAEBLE was validated for Lake Mead in the southwestern United States. In this work, we assess STAEBLE for Lake Suwa, Japan, which differs significantly from Lake Mead in terms of climate, lake size, and surrounding topography and land cover. We ran STAEBLE for the years 2016-2018 with ERA5 reanalysis and MODIS water surface temperature data, and validated the lake evaporation estimates against in-situ flux observations. We also ran STAEBLE with observed meteorology from automated weather stations in the vicinity of the lake. STAEBLE performed very well at Lake Suwa with the ERA5 data, slightly better than at Lake Mead, compared to flux observations at daily to monthly time scales, with normalized mean bias errors (NMBE) of -4% to -5% and correlations of 0.88 to 0.95 at the monthly scale, depending on the model configuration. When driven by weather station data, the overall model performance further improved. We found that good estimates of net radiation and water surface temperature, both of which are much easier to measure than over-water turbulent fluxes, are important for good model performance.

　Machine learning (ML) techniques have been increasingly used to estimate crop yields at scales ranging from on-site to global. Since ML techniques are data-driven approaches, it is empirically known that the performance of a specific ML algorithm depends on the manner in which the training dataset is compiled. However, few studies have quantitatively evaluated the performance. In this study, global rice yields were estimated through a random forest (RF) methodology. Performance dependency of RF on training data was examined by a comparison of estimated yields using different training datasets covering different yield ranges and geographical extents. First, 14 explanatory variables collected from different sources (satellite vegetation, meteorology, and geographical location data) were used for building RF regressors. The crop calendar was determined from a combination of satellite vegetation and crop model simulation. Next, RF regressors were trained to give census-based rice yields (used as reference yields) from training datasets of the 14 explanatory variables. By applying the RF regressors to validation datasets, misfits between estimated and the reference yields were evaluated. RF reproduced rice yields, but the accuracy depended on the training data. Yields beyond the yield range of the training data could not be reproduced by RF. This indicates that the yield range of the training data determined the possible range of estimated yield. Among the 14 variables, geographical coordinates (longitude and latitude) ranked the highest importance, i.e., played a crucial role in estimating yields. The RF regressors built from the 14 variables outperformed those built only from the geographical coordinates in accuracy but with limited advantage. We concluded that (1) choosing training data to cover all possible yield ranges of the target rice-cropping areas was crucial for accurate yield estimation using RF and (2) incorporating satellite and simulation data was advantageous for building high-performance RF regressors.

　Apple scab (Venturia inaequalis (Cooke) Winter) is one of the most devastating diseases of Apple (Malus pumila var. domestica) cultivation. This study focused on the impact of global warming on the occurrence of apple scab and investigated a possible future risk at Aomori Prefecture, Japan. We applied a large-ensemble climate change dataset to the Pear Disease Occurrence Forecast Model and evaluated the infection risk of apple scab. The severe infection risk in the present-climate data was high in spring and after a slight reduction in summer, rose again in autumn, consistent with the priority control period from April to June. Moreover, the probability of severe risk during the priority control period only slightly reduced in future experiment, corresponding to 2090s of a high-end scenario. The risk appearance date defined as the 100-year return period was March 29 in the future, 16 days earlier than in the present climate. The hypothetical experiment revealed that the results can be explained mainly by a temperature increase almost by 5 K from 1980 to 2090. Furthermore, since warmer temperatures exceeded the growth threshold for Venturia inaequalis, the infection risk significantly reduced to the level of 100-year return period in summer in future.

　Field experiment data or crop observations at sites reported in agronomic literature are of high quality and have been considered as a potential source of information for the development of a global grid crop dataset. However, extracting data on a crop variable of interest from the text and tables of many papers is a time-consuming, painstaking task for dataset developers. Recent advances in large language models (LLMs) and resulting tools are expected to provide a promising solution. This study presents a computational method for extracting data from research papers using an LLM-based online tool, ChatPDF. The Python program we developed is applied to the 164 papers to extract crop phenology data of maize, soybean, wheat and rice for demonstration purposes. The results show that the LLM-based data extraction method can dramatically reduce the burden of data extraction in human curation, but needs improvement to become a reliable alternative that can replace manual data extraction. In particular, innovations are needed to increase the capture rate by avoiding data omissions and to reduce errors by correctly inferring longitudes, latitudes and harvesting years. The LLM-based data extraction is currently in its infancy and deserves future research for large-scale implementation.