In this study, two commonly used drought indices; the Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI), were analyzed in order to understand the impacts of drought on maize yield over four main maize production provinces of South Africa. The drought was characterized using three Drought Monitoring Indicators (DMI) i.e., the Drought Duration (DD), Drought Severity (DS), and Consecutive Drought Months (CDM). The results indicate that maize yield is significantly affected by drought across the entire study area, although the impacts are localized. A comparison between the SPI and SPEI with maize yield suggests that the SPEI is more correlated and sensitive to maize yield than the SPI. The maize yield is particularly most sensitive to the 3-month SPEI. The 3-month accumulation period coincides with maize growing season (r = 0.59; p < 0.05). The analyzed results illustrate that drought affects maize yield by up to 35% across the study area. Additionally, results depict inherent spatial patterns of DMIs demonstrating that there are differentiated drought impacts across the maize production areas. The results suggest that management strategies that allow for optimal water use within the first 1- and 3-month periods would be most effective for sustainable maize production within the study area. This research study contributes towards a deeper understanding of the characteristics of drought and their impacts on maize crop production. Such knowledge is important in e.g., the formulation of drought monitoring and prediction strategies including drought early warning systems.
Over a period of 9 years covering both disturbed and undisturbed periods, we evaluated long-term variations in water and light use efficiency (WUE, LUE) in a cool-temperate mixed forest in northern Hokkaido, Japan: our study clarified the relationship of WUE and LUE to environmental and vegetation variables. WUE and LUE markedly decreased after deforestation: they were negatively correlated with the seasonal variations in photosynthetically active radiation (PAR) and water vapor deficit, and LUE increased with the increase in the leaf area index (LAI) coincident with the vegetation recovery. Other environmental and vegetation variables did not explain the changes in the WUE and LUE. The effect of clear-cutting on LUE was well explained by the change in the LAI; on the other hand, vegetation composition (developed forest or young tree seedling with dense undergrowth) had little effect on the magnitude and variation of LUE. In contrast, the change in LAI had little effect on WUE, because WUE is more sensitive to the atmospheric water deficit than it is to vegetation structure.
Evapotranspiration estimates in forested areas are important not only for water resource management on a regional scale but also to better understand the water cycle on a global scale. The objective of this study was to evaluate the Global Change Observation Mission-Climate (GCOM-C) global Evapotranspiration-index (ETindex) estimation algorithm (GCOM-C ETindex algorithm) applied to forested areas. ETindex, which is the ratio of the actual evapotranspiration to the reference evapotranspiration, is estimated from the actual surface temperature and hypothetical wet and dry surface temperatures, i.e., Ts (wet) and Ts (dry), respectively. Based on the algorithm, evapotranspiration is calculated from thermal satellite images and near-surface weather data. We compared the observed ground-based annual evapotranspiration with the estimated annual evapotranspiration obtained using the GCOM-C ETindex algorithm and thermal images from the Moderate Resolution Imaging Sectroradiometer (MODIS) satellite for 40 forests, with 10 sites in four different areas, including Japan, North America, Australia, and the tropical region. We found that the GCOM-C ETindex algorithm well reproduced annual evapotranspiration for most forests. The root mean square errors (RMSE) for the 40 forests was 239 mm. In Japan, North America, and Australia, the overestimation of summer evapotranspiration was offset by the underestimation of winter evapotranspiration. The accuracy of annual evapotranspiration estimates in forests with low annual mean temperatures (<15 °C) was less than that in forests with high annual mean temperatures (≥15 °C). Forests with a low annual mean temperature displayed low levels of evapotranspiration in winter. In these forests, the overestimation of summer evapotranspiration was not offset by the underestimation of winter evapotranspiration. The overestimation of Ts (wet) is the primary reason for the overestimation of summer evapotranspiration. Redetermination of the parameters for the Ts (wet) estimates must improve the evapotranspiration estimates in the forested areas, especially the ones with a low annual mean temperature.
Global warming has already affected agricultural production worldwide. In Japan, the occurrence of chalky rice grains (CRGs), which are low quality grains whose occurrence is exacerbated by high air temperature, has become a major issue. There is concern that this occurrence will become more frequent with increasing global warming; therefore, it is necessary to develop a model that can accurately estimate the impact of global warming on the occurrence of CRGs. In this study, such a model was developed for Japan’s main cultivars, “Koshihikari” and “Akitakomachi”. This model is a non-linear model and was also used to investigate the differences in the responses of these cultivars to temperature and radiation; the occurrence of CRGs was explained as a function of air temperature and solar radiation in this model. The results showed that the occurrence of CRGs could be reproduced more accurately when both air temperature and solar radiation were considered compared to when only air temperature was considered; the latter is the case with existing models. The air temperature at which CRGs begin to appear is lower for “Koshihikari” than “Akitakomachi”, when the average solar radiation is small (less than 16.4 MJ m-2 d-1). The proposed model will hopefully be applied to impact assessments of global warming and the quantification of varietal characteristics with respect to the occurrence of CRGs.
Leaf area is one of the most important elements of information in plant management. Leaf area is associated with many agronomic and physiological processes including growth, photosynthesis, transpiration, photon interception, and energy balance. Three-dimensional (3D) plant architecture is required for monitoring, since plants have three-dimensionally complex structures. A photogrammetric approach called structure from motion (SfM) was used for the 3D measurement. A method using the total area of a horizontal face of voxels could possibly be employed to estimate leaf area in 3D plant images. However, the leaf inclination angle of each small part, the voxel size, and misconfigured voxels in a vertical direction near leaf surfaces should be considered in the calculation. We propose a method for leaf area estimation in voxel-based models that overcomes these problems of estimation error. Using our method, the leaf area was estimated with an absolute error of 8.87 %. This result was obtained by fully utilising 3D information such as voxel size and leaf inclination angle at each voxel. Moreover, our method does not involve manual operations for its construction, unlike a previous method. From the perspectives of high degrees of accuracy and automatic procedures, this voxel-based leaf area calculation method is advantageous.
Several eco-physiological process-based crop models have been used in combination with climate models to predict agricultural yield to assess the impact of climate change. However, the quality degradation of rice caused by the influence of climate change is a prevailing problem. Although there is extensive elucidation of the mechanism of the occurrence of white immature grain because of high temperatures resulting in quality degradation, there are fewer studies that incorporate this into their prediction models. In this study, a statistical model to estimate the first-grade rice ratio was developed for three major rice cultivars in Japan. Parameters for a heat-dose index were estimated by employing the particle swarm optimization method and parameters for the statistical model were estimated with the maximum likelihood method. Parameters of the statistical model varied depending on the cultivar variety. It was observed that the statistical model showed varied prediction accuracy for the first-grade rice ratio based on the temperature that was incorporated into model, that is, daily mean, maximum, or minimum temperatures. Our result can generate more accurate predictions of the impact of climate change on rice production, incorporating the farmers’ choice of adaptation to climate change, including the shift in transplanting day.