Comparison of Length and Transport of Entrapped Woody Debris in Coniferous and Broadleaf Forests Based on Mapping Using Ortho-Photographs Acquired by Uncrewed Aerial Vehicle

Landslides and debris ows often result in woody debris from initiation and riparian zones, through their runout. Considering that woody debris is one of the main components of watershed ecosystems, the importance of quantifying its properties and transport is evident. However, the low accessibility of disturbed channels after landslides and debris ows generally impedes the accurate and quick investigation of woody debris. The recent advances in photogrammetry techniques and technology may overcome such issues. In this study, we used ortho-photographs acquired by a small uncrewed aerial vehicle (UAV) for measuring the lengths of woody debris entrapped mainly by closed-type check-dams. We focused on two channels, located in coniferous and broadleaf forests and affected by two different landslides events. The measurement accuracy was analyzed by comparing the lengths derived from the UAV method with direct measurements. When both edges of woody debris were satisfactorily extracted from an ortho-photograph acquired via UAV, the length of the woody debris with respect to coniferous trees can be measured with an accuracy of approximately ±0.5 m. However, some coniferous trees were captured by stand trees in the riparian zone, and the coverage by tree-crowns led to the underestimation by several meters of the extracted length of the entrapped woody debris. For broadleaf trees, most of the extracted lengths were shorter than the directly measured lengths. This is probably caused by the low visibility of both edges due to the complex structures of the root-wad and the tree-crown. Our results showed that there were no signicant changes in the lengths and locations of the entrapped woody debris, in both sites, after seven months of the rst UAV ight. In the coniferous forests, the rainfall that triggered landslides in 2017 exceeded the 100-year return level, which was obviously an abnormal intense rainfall. Although the 2019 rainfall event that occurred between UAV ights was not as much to the rainfall triggering landslides, rainfall intensities with different durations reached the second-highest value from 1976 to 2019, exceeding the 30-years return period. This suggests that most of the entrapped woody debris rarely migrate even under extreme rainfall.


Introduction
Debris ow causes entrainment of stand-woods located in the initiation and riparian zones, and, consequently, may include dozens of percentage of woody debris by volume (Johnson et al., 2000;May and Gresswell, 2003a;Lancaster et al., 2003). In addition to its destructive impact on life and infrastructure (e.g., Ruiz-Villanueva et al., 2013), woody debris in debris ow can alter the ow regime because of their irregular shape that can be entrapped around obstacles, leading to anomalous deposition of sediment and inundations (e.g., May, 2002;Lancaster et al., 2003;Tang et al., 2018;Booth et al., 2020). Therefore, the accumulation of woody debris results in structural peculiarities around the channel networks (e.g., Keller and Swanson, 1979; Woodsmith and Bu ngton, 1996; Montgomery et al., 1996;Nakamura and Swanson, 2003). This contributes to changes in the ecosystem, the channel morphology, and the sediment ux through woody-debris preservation and decay (e.g., Wallace and Benke, 1984;Lisle, 1995 . Hence, the river form and function are determined by the interaction between water, sediment, and wood (Nakamura et al., 2017;Swanson et al., 2020). Quantifying woody debris is important for assessing its impacts on ecological, geomorphological, and uvial conditions. Many previous studies focusing on in-situ channels have contributed to unraveling the role of woody debris in various spatiotemporal-scales and environmental settings, but most of these approaches required direct eld measurements, such as local monitoring (e.g., Manners et al., 2007), tracking of woody debris (e.g., Ravazzolo et al., 2015;Wyżga et al., 2017), and eld experiments using arti cial woody debris (e.g., Haga et al., 2002). Taking into account the diversity of forests around channels (e.g., age, species, and density of trees), the accumulation of eld data is obviously an effective approach.
However, the necessity of human effort in the eld hinders data acquisition at inaccessible areas (e.g., headwater channels and disturbed areas immediately after landslides and debris ows). In practice, the eld data acquisition is di cult for a large scale area exceeding a sub-basin size. The improvement of eld measurement techniques with respect to woody debris remains a critical issue.
To address the measurement issues of woody debris, remote sensing approaches using threedimensional data have been applied. The Use of LiDAR (light detection and ranging) data clearly reduces the processing time required for mapping the logjam and large woody debris (Kasprak et al., 2012;Abalharth et al., 2015;Atha and Dietrich, 2016). However, such technology is expensive and therefore available only in some regions. Alternatively, photogrammetry based on structure from motion multi-view stereo (SfM-MVS) using UAV (uncrewed aerial vehicle) has been proven time-e cient compared to classical eld surveys (Sanhueza et al., 2019). This approach overcomes data availability issues and is relatively low-cost. Nevertheless, most tests were conducted in low-land and ood plains rather than lowaccessibility areas such as steep channels (e.g., Sanhueza et al., 2019). As the accuracy of SfM-MVS is remarkably in uenced by complex surfaces and obstacles, such as steep slopes, large reliefs, and vegetation coverage (e.g., Fonstad et al., 2013;James and Robson, 2014), many unresolved uncertainties remain over the application of the SfM-MVS approach in steep and complex targets, such as woody debris in channels.
Aerial photography being one of the traditional two-dimensional data sources may more or less provide meaningful information on woody debris. Even satellite images from Google Earth are being used as accurate for mapping woody debris (Atha, 2013;Ulloa et al., 2015). Hence, the e cacy of aerial photography is evident, but the accuracy and effort of mapping depend on the image quality. In this respect, it is expected that small UAVs allow acquiring high-resolution aerial photographs at low cost because of the lower ight-altitude and higher portability compared to conventional aerial vehicles. Moreover, because ights of small UAVs can overcome inaccessibility issues and cover several kilometers depending on the ight design, it is a fairly attractive tool for obtaining woody debris measurements in low-accessibility areas.
In addition, although entrapped woody debris is often quickly removed to avoid unexpected damages in the downstream area due to its transport, UAV ights can be carried out immediately after rainfall events that result in a large amount of woody debris. Conversely, the risk of secondary impacts arising from woody debris entrapment has not been evaluated properly so far, because it has been di cult to conduct eld surveys immediately after intense rainfall involving woody debris supply. Therefore, mapping woody debris would enable investigating the possibility of woody debris transport due to subsequent rainfall, even if it is carried out based on a simple method using aerial photography. Nevertheless, the potential to measure woody debris based on aerial photographs acquired via UAV has not been thoroughly examined due to lack of sample cases.
In this study, we analyze the accuracy of UAV-based measurements in the case of entrapped woody debris. Two regions are selected representing two forest types, coniferous and broadleaf forests. In both regions, large amounts of woody debris were supplied through landslides and debris ows triggered by a single rainfall event. This research has two main objectives: (1) to analyze the capability of orthophotographs acquired via UAV to measure the lengths of entrapped woody debris, and (2) to investigate the transport potential of entrapped woody debris based on rainfall analysis. Based on the results, we discuss the effectiveness of UAV measurements and how woody debris behave after their entrapment.

Study Site
To examine the in uence of tree type on the measurement accuracy of woody debris using the orthophotographs acquired by UAV, two different sites covered by coniferous or broadleaf forests were selected as study sites (Fig. 1a). Hereafter, the former site is referred to as the CF (coniferous forests) site (Fig. 1b), while the latter site is referred to as the BF (broadleaf forests) site (Fig. 1c).

Coniferous forests site
The CF site is a sub-watershed of the Otoishi watershed located in the northern part of the Kyusyu Island, Japan ( Fig. 1a, b). Its highest point is about 385 m a.s.l., and its lowest point is about 225 m a.s.l. The total length of the main channel is approximately 760 m with an average slope of around 12° (before the debris ow occurrence), and the drainage area is about 0.15 km 2 . To prevent hillslopes and banks destabilization, ve closed-type check-dams were constructed before the debris ow occurred in 2017 (Fig. 2a). Most hillslopes are covered with arti cial coniferous forests that consist mainly of Cryptomeria japonica and Chamaecyparis obtusa. The tree height ranges from about 15 to 25 m.
In the region around the CF site, the intense rainfall on July 5, 2017, triggered more than 2,000 landslides, which resulted in debris ow propagations and therefore a great amount of woody debris (Chigira et al., 2018). The sliding sediment layers were mainly granodiorite and pelitic schist (Chigira et al., 2018). The intense rainfall in July 2017 triggered seven landslides at the CF site (Fig. 2a), but the existing checkdams effectively prevented the descent of sediment and woody debris. Hence, most of the woody debris was trapped around the check-dams and along the channel, preventing the damages in the downstream residential area. Because of this low impact on the residences, woody debris had been maintained approximately untouched until the time we carried out observations.

Broadleaf forests site
The BF site is located in the Hiroshima prefecture, the west part of Japan (Fig. 1c). Here, two channels are con uent at an irrigation pond that is located at the lower end of channels (~ 305 m a.s.l., Fig. 2b). The highest points of channels A and B are ~ 410 and ~ 480 m a.s.l., respectively. Their total lengths are ~ 880 and ~ 466 m, respectively. The average slope of both channels is similar at ~ 11-12° (before the debris ow occurrence). The drainage areas of channels A and B are ~ 0.07 and ~ 0.13 km 2 , respectively. Similar to the case of the CF site, four closed-type check-dams have been constructed before the debris ow occurred in 2018 (Fig. 2b). The forest type is completely different, and most of the hillslopes are covered by broadleaf forests with various tree types. The tree height ranges from about 10 to 20 m.
In the Hiroshima prefecture, the stationary front affected by Typhoon Prapiroon caused heavy rainfall on July 5-7, 2018 (Tsuguti et al., 2019) and triggered approximately 8,000 landslides due to the vulnerable geological setting mainly dominated by weathered granite (Kaibori et al., 2018). Five landslides occurred at the BF site, but the check-dams and the pond effectively prevented the descent of sediment and woody debris. Similar to the situation of the CF site after the disaster, the woody debris produced in the BF site had been almost remained untouched until we carried out the observations.

Methods
This study consists of two parts: (1) accuracy assessment of the woody debris measurement by orthophotography acquired via UAV, and (2) rainfall analysis to investigate the possible transport of entrapped woody debris.

UAV ights and aerial-photograph processing
Two UAV ights were conducted in the study sites (Table 1). A small UAV (DJI Mavic 2 pro; Table 2) was used for the ights, covering the channel and landslides (Fig. 2). The ights were conducted manually at elevations between about 50 m and 200 m, and most of the photographs were taken in nadir direction.
The ight path accounted for an overlap in the aerial-photographs of at least 70%. To reduce differences in shade and brightness among the acquired aerial-photographs, the ights were operated by avoiding the strong sunlight.  The acquired photographs were processed using the SfM-MVS photogrammetry software (Agisoft, Metashape Professional version 1.5.1). In most cases, immediately after debris ow disaster, woody debris generally impede the setting of ground control points (GCPs) as they behave as obstacles. Considering this, our intention was to assess the accuracy of measurements using only UAV ights (without any other eld operations). Hence, the position data of a built-in GNSS (global navigation satellite system) contained in exchangeable image le format data of each photograph was used for georeferencing. Following the standard SfM-MVS photogrammetry work ow, without geo-referencing using GCPs, the ortho-photographs were produced with a spatial resolution of 0.03-0.05 m.
Because the processing was performed without georeferencing by GCPs, the coordinates of the resulted ortho-photographs taken during the different ight dates did not overlap. As the aim of using sequential ortho-photographs is to investigate the possible migration of the entrapped woody debris, it is necessary that the relative coordinate system overlap, even if the absolute coordinate system did not. Thus, we manually adjusted two sequential ortho-photographs by using the crests of the check-dams as reference points, whose coordinates were taken from the older ortho-photograph. Using the extracted coordinates, the ortho-photograph acquired on the later date was transformed to t the rst one. The differences in lengths regarding selected the immobilized objects (e.g., large boulders and crests of check-dams) were less than about 0.2 m. Therefore, the horizontal accuracy of ortho-photographs was considered as less than a few decimeters.

Accuracy assessment
First, the length of entrapped woody debris was extracted as line data from ortho-photographs. The measurement accuracy of the extracted length is presumably affected by visibility, which is related to the types and sizes of the trees. Moreover, because woody debris is often inclined and broadleaf trees have a curved shape, the orthogonal projection in the extracted length possibly causes a measurement error ( Fig. 3). Hence, we focused on the differences due to tree types, diameter, and entrapment inclination of the woody debris. In the selected woody debris, the length excluding the root-wad and twigs was directly measured and compared with the extracted length (Fig. 3). The diameters of the woody debris were measured at about 1.0 to 1.5 m height from the root-wad. For small woody debris of less than about 1.5 m, because the difference in the thickness of a single fragment was slight, we measured the diameter at an arbitrary location. In the CF site, as woody debris remained relatively straight shape, the entrapment inclination was directly measured (Fig. 3b). The results of direct measurements were compared with the length extracted from the ortho-photographs of the rst UAV ight.
The extraction of woody debris depends on its visibility in the ortho-photographs, and thus the identi cation of its origin as well as the detection of small woody debris would go beyond the measurement purpose. Moreover, the extracted woody debris may contain multiple fragments that originated from a single tree. In other words, as a single stand-wood may result in several batches of woody-debris, counting the extracted woody debris involves uncertainty. Considering this unavoidable risk of over-or under-estimation in the amount of woody-debris, we avoided a detailed interpretation and analysis using the extracted numbers of woody-debris.
Analysis to determine the transport of woody debris To analyze whether or not woody debris migrated through time, we created quantile-quantile (Q-Q) plots with respect to the probability density of the extracted length. The CF site was divided into seven regions, six landslide scars and the riparian zone (Fig. 2a). Accordingly, changes in the probability density of the woody debris length were visually tested using the Q-Q plots. Similarly, Q-Q plots were made for the channels A and B of the BF site (Fig. 2b). For each region, to investigate the coincidence of the probability density of the extracted length between the two ight dates, a goodness-of-t test was carried out using the Kolmogorov-Smirnov test.
In addition, we investigated whether speci c rainfall characteristics triggered the secondary transport of woody debris. For this reason, we used the hourly rainfall records from 1976 to 2019 obtained at the closest rain-gauges installed by the Japan Meteorological Agency, the Asakura station (for the CF site, Fig. 1b) and Higashihiroshima station (for the BF site, Fig. 1c). The annual maxima of rainfall intensity with various durations (1, 2, 3, 6, 12, 24, 48, and 72 hours) were investigated from 1976 to 2019 and compared with the intensity of the rainfall that triggered the debris ow. In addition, we examined records of rainfall maxima after the debris-ow occurrence. The return periods were calculated based on the probability density of each rainfall intensity tted by the Gumbel distribution based on the annual maxima (e.g., Koutsoyiannis et al., 1998;Sane et al., 2018). For this tting process, the goodness-of-t test (Kolmogorov-Smirnov test) resulted in p-values exceeding 0.1, suggesting that the applied Gumbel distribution models tted the probability density of the observed rainfall well.

Results
Length of the entrapped woody debris As expected, the comparison between the extracted and measured length of the woody debris indicates that the measurement accuracy depends on the visibility of the woody debris entrapments. Fifty batches of woody debris were directly measured at the CF site (Fig. 4a). Fourteen of them were trapped by standwoods located in the riparian zone, and they were partially invisible from the sky due to the coverage by the tree crown. This low visibility caused underestimation of the extracted length, with errors in length at a maximum of ~ 10 m. By contrast, 36 other batches were fully visible, and their lengths ranged between about 3 m and 22 m. Despite this wide range, most measurements coincided well with the extracted lengths. This clear dependency on the visibility of woody debris was re ected in the box-plot (Fig. 5), indicating that the interquartile ranges obviously differed between visible and covered woody debris; the differences were from − 0.4 to 0.5 m and from 1.6 to 4.3 m, respectively (Fig. 5).
The length of 130 woody debris batches without crown coverage was directly measured at the BF site but rarely coincided with the extracted length, which was usually an underestimation (Fig. 4b). Moreover, the differences between the lengths varied. The related interquartile range ranged between about 0.3 m and 2.1 m, re ecting low accuracy and high uncertainty in the measurement of broadleaf trees compared to coniferous trees (Fig. 5).
Strictly speaking, the extracted length was the orthogonally projected length, and therefore, we calculated the slope distance of the measured woody debris from the extracted length using the measured slope. However, this correction did not signi cantly improve the measurement accuracy because differences in the visibility of the woody debris and the tree types overwhelmed the in uence arising from the approximation by the orthogonal projection (Figs. 4a and 6). The differences in the diameter of the woody debris indicate that the extracted length of relatively narrow batches ( < ~ 0.2-0.3 m in diameter) tends to be underestimated (Fig. 7), suggesting that it depends on visibility.
At the CF site, the interquartile ranges slightly differ among the landslide scars, but they are similar between the riparian zone and the average of all landslides, which is 2.5 to 7.7 m and 2.6 to 8.3 m, respectively (Fig. 8a). The length median of the woody debris ranged from 3.4 m (in landslide 4) to 7.8 m (in landslide 1). At the BF site, the interquartile ranges of the length of the woody debris at the checkdams A and B were 1.1 to 3.1 m and 0.8 to 2.6 m, respectively (Fig. 8b). The medians of the check-dams A and B were similar, 1.8 m and 1.5 m, respectively. Consequently, the length median of the woody debris at the CF site was ~ 4.5 m and obviously greater than that at the BF site (Fig. 8c).

Transport of entrapped woody debris
At both sites, the visible changes in the spatial distribution of the woody debris between the two orthophotographs were slight (Figs. 9 and 10). Woody debris was trapped in various areas, for instance, as stand-woods in the riparian zone, the check-dams, and the landslide areas. Despite these differences in the entrapment regions, most of the woody debris seems did not move even after about 7 months (at the CF site, Fig. 9) and over one year (at the BF site, Fig. 10). The Q-Q plots indicate that the probability density of the extracted woody debris length coincided well with the measured length, regardless of the division of the entrapment regions (Fig. 11). Thus, the p-values with respect to the goodness-of-t test exceeded 0.05, suggesting that the changes were statistically small.
The rainfall anomalies that triggered debris ows were revealed by comparing the annual maxima of the rainfall intensity in each case. At the CF site, the rainfall that triggered debris ow in 2017 was obviously excessive rainfall, as the rainfall intensity was the highest compared with the annual maxima of other years, regardless of the rainfall duration (Fig. 12a). This suggests that the highly intense and continuous rainfall resulted in the large amount of woody debris. At the BF site, regarding the rainfall that triggered debris ows in 2018, rainfall intensities with 1-6 hours durations were not higher than the other annual maxima since 1976, but those with 12-72 hours durations were the most intense since 1976 (Fig. 12b). This suggests that the high rainfall intensity continuing over 12-hours resulted in landslides and debris ow with the propagation of woody debris. Both rainfall events that triggered debris ows at the CF and BF sites reached the level exceeding the 100-years return period (Tables 3 and 4).  At the BF site, after the landslides and debris ow in 2018, the annual rainfall intensity maxima in 2019 was not remarkable (Fig. 12b, Table 4). Thus, it is di cult to analyze the potential of secondary transport of entrapped woody debris via intense rainfall. By contrast, at the CF site, after the landslides and debris ow in 2017, the annual rainfall intensity maxima in 2018 and 2019 were the second or third highest values since 1976 (Fig. 12a). Although the intensity of the triggering rainfall in 2017 signi cantly exceeded the annual maxima in 2018 and 2019 in terms of the return period, some rainfall intensities in 2018 and 2019 exceeded the level of the 30-years return period (Table 3). Such a heavy rainfall event occurred for example on July 21, 2019 (in between the UAV ights at the CF site; Table 2).

Implications for UAV-based woody debris mapping
Our results demonstrate that the visibility and tree types of the woody debris determine the ability to accurately measure the length of the woody debris using ortho-photographs acquired via UAV. The length of most of the woody debris batches originating from coniferous trees was measured with less than ± 0.5 m accuracy using the ortho-photograph acquired by UAV, whereas in the case of broadleaf trees the length was underestimated systematically despite the absence of tree-crown coverage lowering visibility (Figs. 4b and 5). Considering the complex structures of both ends of broadleaf trees (i.e., wide and rounded crown and root-wad; Fig. 3b), this underestimation may be attributed to missing extractions due to low visibility of the broadleaf trees arising from complex structures at the ends. In addition, the length of the entrapped woody debris at the BF site was less than that at the CF site (Fig. 8), implying the fragmentation of broadleaf trees during the debris ow. Although the characteristics of debris ow also probably in uence the magnitude of woody debris fragmentation (e.g., Johnson et al., 2000;Lancaster et al., 2003), the fragility of broadleaf trees may reduce the measurement accuracy based on the orthophotograph at the BF site by decreasing the visibility of woody debris.
Hence, most of the factors that decrease the measurement accuracy of woody debris depend on visibility. Based on this interpretation, woody debris measurement via UAV has both positive and negative aspects. Several components affecting the quality of the aerial photographs (e.g., black shadow, brightness involving overexposure, and overlap ratio among aerial photographs) depend on the weather conditions and the ight plan (e.g., path, height, and direction of shooting). An adequate ight plan may signi cantly improve the accuracy of UAV-based woody debris measurements. Moreover, the recently developed highaccuracy built-in GNSS-based direct geo-referencing in the SfM-MVS photogrammetry processing makes it possible to add accurate absolute coordinates without GCPs (e.g., Carbonneau and Dietrich, 2017). As part of our analysis includes manually adjusting the coordinates, this improvement is also expected to increase the measurement accuracy.
In addition, our results reveal an obvious limitation of ortho-photograph-based measurements. In general, woody debris is not only dispersedly distributed along a channel, but also results in logjam due to stacking and accumulating (e.g., Manners et al., 2007;Abalharth et al., 2015). However, the correlation between measurement accuracy and visibility from the sky demonstrates that quantitative analysis of logjam is beyond the scope of ortho-photograph-based measurements. Consequently, the total amount of woody debris is probably underestimated in all cases.
Possibility of the transport of the entrapped woody debris The slight changes in the length and spatial distribution of the woody-debris suggest that the woody debris batches, entrapped by the check dams, did not move after the debris ow disaster (Figs. 9, 10, and 11). The elapsed time from the rst ight was about seven months for the CF site and over a year for the BF site (Table 2), which may be insu cient to cause signi cant visible decay. This means that intact woody debris still remains in the upper regions of the channels. However, considering the slight differences in the spatial distribution and the Q-Q plot of the woody debris at the CF site (Figs. 9 and 11), the extremely high rainfall intensities in 2019 ( Fig. 12a and Table 3) suggest that most of the entrapped woody debris remain in place and rarely migrate, even when intense rainfall exceeding the level of the 30years return period occurs. This implies that most of the rainfall events that do not cause debris ows and landslides cannot trigger secondary transport of entrapped woody debris.
Because the pre-existing check-dams trapped woody debris at both sites, differences in the trend between such arti cial trapping and natural entrapment of woody debris should be noted. Even so, the low transportability of the entrapped woody debris provides insight into the role of the supplied woody debris and sediment in the upstream channel after debris ows and landslides. The e cacy of woody debris trapping by closed-type check-dams is generally considered relatively low compared with the case of open-type check-dams (e.g., Piton and Recking, 2016). Nevertheless, the e cacy of woody debris trapping by closed-type check-dams was gained with long-term due to the low possibility of secondary transport. As a consequence, this low transportability of entrapped woody debris implies low erosion and discharge rate of the stored sediment, indicating that woody debris can increase sediment storage in headwater channels (e.g., May and Gresswell, 2003b). This therefore suggests that the trapping woody debris and sediment is most important for preventing their descent.

Conclusions
We mapped the entrapped woody debris using ortho-photographs acquired by UAV in two headwater channels covered by coniferous or broadleaf forests to: (1) determine whether or not ortho-photography acquired vis UAV allows measuring the length of entrapped woody debris, and (2) investigate the possibility of transport of the entrapped woody debris by intense rainfall.
At the CF site, comparing the directly measured length with the length extracted by the UAV mapping indicates that aerial photography measurements have approximately ± 0.5 m accuracy, when both edges of the woody debris batch were satisfactorily con rmed. In contrast, some woody debris batches were trapped by stand-woods in the riparian zone, which were invisible underneath the tree crowns, and this led to underestimation of the extracted length. At the BF site, most of the extracted length values are shorter than the length measured directly, probably due to low visibility of the batch edges due to complex tree structures in the root-wad and the tree crown. Hence, these results suggest that the measurement accuracy of woody debris via UAV strongly depends on the visibility of the woody debris arising from tree types.
At both sites, changes in the frequency distribution of the length and the location with respect to entrapped woody debris were slight even over seven months had passed since the rst mapping.
Nevertheless, at the CF site, some rainfall intensities between the two ights reached the second-highest value since 1976. Although the rainfall intensity that triggered landslides in 2017 was higher than the annual maxima in 2019, some intensity values in 2019 exceeded the 30-years return period. This suggests that most of the entrapped woody debris remain and rarely migrate again, even under intense rainfall. The data used in this paper are available from the authors upon request.

Competing interests
The authors declare that they have no competing interest.

Funding
This work was supported by the project "Development of technology for impacts, mitigation and adaptation to climate change in the sectors of agriculture, forestry and sheries" of the Agriculture, Forestry and Fisheries Research Council.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.