Understanding the historical processes of biodiversity degradation is essential for establishing goals and objectives for ecosystem restoration. In this study, records of local extinction of freshwater fish were collected from national and prefectural Red Data Books (RDBs), and the processes of biodiversity loss in Japanese local regions were investigated, taking into account the ecological and sociological backgrounds of inland water ecosystems specifically before the 1970s. The descriptions in RDBs were scrutinized, and information that directly or strongly indicated local extinction was extracted for each species or taxon. Additionally, causes and timings of local extinction were collected, if available, and compiled by decade. A total of 162 cases indicating local extinction were collected, and timing of local extinction was determined for 84.0％ of all cases, with 55.9％ of these last records occurring before the 1970s. Habitat degradation was the largest cause of local extinction (46.5％), followed by competition with and/or predation by exotic alien species (12.7％) and fragmentation from the sea (7.9％). Habitat degradation accounted for most of the causes of local extinction before the 1930s, fragmentation from the sea was 25.0％ in 1940, and competition and/or hybridization with closely-related alien species was 23.1％ in the 1950s. The causes of local extinction diversified after the 1960s. The use of pesticides and intensive catch for ornamental use increased from the 1960s to the 1970s and from the 1960s to the 2000s, respectively. Local extinction due to the invasion of exotic alien species increased after the 1960s and accounted for 21.9％ of the total cases from 1990 and later. The number of cases of local extinction peaked in the 1970s during Japanʼs high economic growth period, indicating explicit biodiversity losses of freshwater fish in this period. Although the rate of local extinction seems to have slowed after the 1980s, new concerns for biodiversity continue to arise. Therefore, implementing intensive measures for the maintenance and restoration of biodiversity is an urgent issue for achieving sustainability.

A treetop detection method was developed to estimate the number and volume of trees in riverine broadleaf forests using airborne laser scanning data. The enhancement in the accuracy of treetop detection was investigated by examining the efficacy of raster data and parameter window size (WS) of local maximum filtering (LMF). A digital canopy ruggedness model (DCRM) was created to represent the broadleaf tree canopy shape characteristics in the raster data. The effectiveness of the model was compared with that of a commonly used digital canopy height model (DCHM). For this study, we employed estimation equations based on tree height (DCHM) proposed in prior research and Fitting WS, an estimation equation fitted via observation data, to evaluate the treetop detection accuracy. The combination of DCRM and Fitting WS yielded the highest F-score (0.62) for treetop detection. The study areas dominated by hackberry trees showed an F-score of approximately 0.70, whereas the willow-dominated area showed a low F-score of 0.41. The obtained accuracy was higher than that of DCHM or other WS cases. This finding suggested that DCRM and Fitting WS are more suitable for areas dominated by riverine broadleaf forests and that LMF-based treetop detection could yield lower accuracy in riverine forests dominated by willows. The combination of DCRM and Fitting WS yielded a reasonably accurate estimate of the number of trees, with an error of -2％. However, the estimation accuracy of tree volume was lower compared with the number estimation accuracy, with an error of -30％. This difference in accuracies could be attributed to the low estimation of tree height, however, the observed tree heights may contain errors in field measurements. Therefore, cautious interpretation of the tree height and volume accuracies estimated with airborne laser scanning data is important.

Unmanned aerial vehicles (UAV) are being used in research about river environments. Users need to be aware of the most suitable conditions in which to photograph a target species and need to develop an efficient method for analyzing the acquired images. In this study, we investigated a method for understanding the distribution of invasive species Solidago altissima using a UAV. We conducted a UAV survey in autumn, when this species is in full bloom, to determine the suitable flight altitude for UAV survey, and to compare the UAV based results with the results of an on-ground vegetation survey. We found that the abundance and distribution of this species could be effectively monitored, using the UAV, because of its conspicuous flowers. In addition, we were able to create a distribution map with high accuracy using deep learning. The results indicate that the method can be applied, as part of river management, to monitor this species.

We revealed the habitat status of the fishes in the upper reaches of the Nogawa River in Tokyo using environmental DNA surveys and sampling surveys. These fishes consisted mainly of those living in the river's middle reaches. In environmental DNA analysis, Opsariichthys platypus, Pseudorasbora parva, and Gnathopogon elongatus showed high read counts. It was also found that some confirmed fish species were endangered and non-native species. The results of the two surveys suggested that the environmental DNA survey could complement the sampling survey. Through these surveys and results, we organized and compared the characteristics of both surveys. The environmental DNA survey and sampling survey were found to have advantages and disadvantages in the "sample collection" and "analysis" stages. The biological monitoring survey considered that using both surveys could provide a more accurate picture of inhabiting fishes and invasive alien species invasions. For resource management, ecosystem management, and conservation, more efficient and accurate results will be obtained by combining multiple survey and analysis methods and considering detailed condition settings, depending on the objectives.