To examine the influence of global warming including increased heavy snowfall frequency on the potential of natural dry snow avalanche frequency and magnitude, we estimated the frequency of weak layer formation and the associated slab overload above the weak layer over northern Japan. The estimation was numerically performed using climate models' output for 1800 winter simulations in each of the historical (1951–2010) and +4°C experiments by forcing a physical-based snowpack model with the result of the climate models. Here the +4°C experiment was defined as a climate when the global mean air temperature had increased by 4°C from the preindustrial level. The estimation results showed that the probability of weak layer formation, identified by the natural stability index, would decrease all over the area because of the shorter age of the weak layers caused by a warmer climate, indirectly indicating a potential decrease in avalanche frequency. However, because of increased heavy snowfall frequency, slab overload would increase by 10–15% in inland areas for weak layers of decomposing fragments/precipitation particles and the mountainous area facing the Sea of Japan for weak layers of facets/depth hoar, thereby potentially indicating an increased magnitude of avalanches.

Although the disaster reduction effects of forest braking have long been known empirically, they have not been known in detail down to recent. In this study, we ascertained forest braking effect by numerical simulations using the avalanche dynamics program, TITAN2D, to model large-scale avalanches. One of these avalanches occurred in the Makunosawa valley, Myoko, and damaged a cedar forest; the others occurred on Mt. Iwate and damaged a subalpine forest. All avalanches damaged many trees and terminated within the forests. In our simulations, the resistance of the forests to avalanches is simulated using a larger bed friction angle. Fitting the observations from the Makunosawa avalanche by trial and error, a bed friction angle of 13–14° in the non-forested area and of 25° in the forested area is obtained. We conducted simulations of the Mt. Iwate avalanches using the same method as for the Makunosawa valley avalanche, and obtained good agreement between observations and simulations. Simulations reveal that without the forest, the avalanche would have traveled at least 200 m farther than the forest's actual end in the Makunosawa valley, and at least 200 m and possibly up to 600 m farther on Mt. Iwate. This study therefore clearly shows that forests provide a braking effect for avalanches.

Snowmelt was measured on a daily basis for 17 days at the open site and 18 days at three Japanese cedar sites with canopy closure of 17.8% (cedar stand A), 5.2% (B) and 2.4% (C) in April. Measured daily snowmelt at each site was reproduced by heat-balance calculation with an accuracy of <±1 mm w.e. From 1st April to the date of snow disappearance net radiation accounted for 88.4, 43.0, 32.7 and 34.2% of total snowmelt energy at the open site, the cedar stands A, B and C, respectively. The ratio of sensible and latent heat to total snowmelt was 33.1–37.9 and 25.9–29.4%, respectively, at three cedar stands. The ratios of sensible and latent heat increased over time in accordance with the rise in temperature at all cedar sites. They became large on a daily basis when air temperature and/or wind speed were high. Wind speed is dependent on morphology around each site that also dictated snowmelt.

The Makunosawa valley, Myoko, Japan, experiences frequent avalanches and is therefore ideally suited to study how meteorological elements influence avalanche activity. Since 2000, five large-scale snow avalanches with running distances >2000 m have been observed and some characteristics of these avalanches have been obtained. However, the characteristics of the snowpack in the starting zones could not be observed because they are too difficult to approach and no snow-pit observations have been carried out. We simulated the variations in the snowpack in the starting zone using the numerical snowpack model SNOWPACK with local meteorological data. The results indicate a layer of faceted crystals with low shear strength followed by rapid loading from snowfall was the cause of three avalanches in February. Conversely, no layer of faceted crystals was shown by the model before a January avalanche and we assume the sliding surface of the avalanche to be precipitation particles. The only wet-snow avalanche is attributed to a decrease in shear strength due to infiltration of meltwater and an increase in liquid water content in the boundary of two layers of different grain sizes.

Accurate measurements of snow avalanche flows in forests are rare. To understand how forests can stop avalanches, we study a mixed flowing avalanche that stopped in a cedar (Cryptomeria japonica) forest in Japan. The large-scale dry slab avalanche occurred in the Makunosawa valley in Myoko on 17 February 2008 and damaged many trees. The site contains a geophone and other devices to detect avalanche activity. A nearby meteorological station provides weather information. Post-event observations were made of the scale of the avalanche and tree damage. These showed that the avalanche released from the east-southeast-facing slope at 1700 ma.s.l. Snow debris reached ~770ma.s.l; the horizontal runout distance was ~3000 m. The area of the runout zone was ~10 ha, and the snow water equivalent of debris was 400–1500 mm. The mass of the avalanche was roughly estimated at 5–10×107 kg (1–2×105m3). We used snowpack simulation models to estimate the height of the fracture zone. The simulations revealed that a faceted snow layer formed at 2 m and the stability index was 0.5 (poor) at the time of the avalanche release, due to heavy snowfall. We estimated the avalanche velocity from the bending stress of the broken cedar trees, in order to verify the effect of forests in reducing avalanche velocity. The avalanche was estimated to flow into the forest at a velocity of >26–31ms–1, with a 17 m high powder snow layer and a dense-flow layer <2m in height. It appears to have slowed down and stopped at ~130m after penetrating the upper edge of the forest.

Long-term data of winter air temperature and precipitation were analyzed and the correlation between them investigated in order to identify the factors influencing snow reduction during the recent warmer winters in the heavy-snowfall areas in Japan. A high negative correlation between winter precipitation and air temperature was identified in the heavy-snowfall areas on the Sea of Japan side in the center of the main island (Honshu). It was confirmed that precipitation is mainly caused by cold winter monsoons, and thus correlates to a large extent with air temperature in these areas. The precipitation decrease can be considered an effective factor for the recent reduction in snow as well as the snowfall to precipitation ratio. This should be taken into account for a better prediction of snow reduction in relation to global warming.

Snowdrifting processes and the wind-velocity profiles around a collector and a blower snow fence were investigated in a cold wind tunnel. The purpose was to ascertain the effect of wind direction on drift control by snow fences. Three different cases were studied for both types of snow fence, and the resultant snowdrifts were compared. In the first case, the snow fence was perpendicular to the wind direction. In the second and third cases, it was tilted by 30° and 45°. When the collector snow fence was tilted, the amounts of snowdrift were much less than when the fence was perpendicular to the wind direction, because the area with low wind velocity was reduced to half behind the tilted fence. On the other hand, the blowing effect of the blower snow fence increased when it was set up at an angle to the wind direction. It is necessary to investigate the position where the blown snow is deposited by the tilted blower snow fence.

In order to facilitate the measurement of liquid-water content of snow in high mountains, a portable calorimeter named “Endo-type snow-water content meter” was developed. It is composed of a metal-coated container made of insulating materials and a lid of the container with a small-thermistor thermometer. Its strong points are its light weight, small size and easy fabrication with cheap materials. The total weight of the device is as light as 250 g, which is less than 10% of the snow-water content meter widely used in Japan (Akitaya-type snow-water content meter). The results of experiments have revealed that the device is capable of measuring the liquid-water content within 2 minutes with an accuracy of 2% by weight.

Disitributions of snow hardness minutely measured with a bandy-type digital load-gauge (push–pull gauge) are demonstrated. This push-pull gauge is of compact design and is useful for precise measurement of tension and compression loads. It can measure the maximum strength of a snowpack when it is destroyed by the attachment pushed horizontally into the side of the pit. Because it takes only a few seconds for one measurement, snow-hardness distribution can be measured at very small space intervals more quickly and with less effort than by using any previous hardness meter, such as a rammsonde, Canadian gauge, Kinosita-type hardness meter, and so on.

Snow-pit observations were made at Saiho, Sapporo and Minakami, Japan, in the winter seasons of 1996 and 1997. The snow hardness was measured with the push-pull gauge at regular intervals of 5 cm vertically and 10 cm horizontally. Some weak layers between harder layers could be detected with the push-pull gauge whereas they could not be using the rammsonde. The hardness of snow was observed to be almost uniform horizontally before snowmelt. Once meltwater infiltrated into the snowpack, its distribution became heterogeneous. It was revealed that the hardness of the fine-grained compacted snow layer with grain-sizes less than 0.5 mm showed a high correlation with the fourth power of the snow density.

Glaciar Upsala, a freshwater calving glacier in southern Patagonia, has been retreating since 1978, and after a drastic recession of about 700 m a−1 in 1994 the retreat seems to have stopped in 1995. A large ice-thinning rate of 11 m a−1 was obtained between 1990 and 1993, by surveying surface elevations near the terminus of Glaciar Upsala. In 1993–94, the thinning was estimated at about 20 m a−1 near the lateral margin. Some possible causes of the thinning behavior are considered.

In the ablation area of Glaciar Perito Moreno, 50 km south of Glaciar Upsala, ablation rates were measured during 110 d in summer 1993–94, and air temperature was continuously recorded throughout 1994. Using a degree-day method with temperature data at the nearest meteorological station, Calafate, annual ablation during the last 30 years was estimated to fluctuate from about 12 ± 2 to 16 ± 2 m a−1 in ice thickness, with a mean of 14 ± 2 m a−1. Thus, the temperature anomaly alone cannot elucidate the thinning of 11 m a−1 at Glaciar Upsala. As a possible mechanism of the ice-thinning, it is suggested that the considerable retreat due to calving may have resulted in reduction of longitudinal compressive stress exerted from bedrock rises and islands near the glacier front, causing a considerable decrease in the emergence flow. Thus, the ice may have thinned at a rate close to the annual ablation rate.