Larix

For the assembly of the L. sibirica genome, four PE and three MP libraries with different insert size were used (Fig. 7 and Additional file 3: Table S3). In the first step, MPE libraries were decoupled and used as single reads to complete a pool of reads. The pool of reads was split to four parts and four sets of contigs were obtained, respectively. The CLC Assembly Cell software was selected for assembling the larch genome as the best performing software.Fig. 7

Sequence coverage for seven sequencing libraries used for the Larix sibirica genome assembly

Also, a fifth set of reads was added to the analysis. This set included all reads, but the PE and MPE reads were decoupled and used as single reads. This set was generated because we found experimentally that the CLC Assembly Cell assembler was able to process the entire volume of the L. sibirica sequence data, but only if the information about the length of the insertion was not indicated. In this case the “Optimization of the graph using paired reads” step is skipped. In this step long repeats are allowed, and scaffolding is not performed which turns out to be too much computationally intense and practically prohibitive for large volume data. Therefore, this set increased the representation of all reads, but they all could be used only as the single end reads at this step.
Unlike the inbred highly homozygous plant used for the genome sequencing and assembly, such as A. thaliana, the L. sibirica tree used for genome sequencing in our study represented a common forest tree with a relatively high level of individual heterozygosity and, respectively, high within individual biallelic variation. The number of ambiguous positions in the L. sibirica sequencing data was estimated at the level of 3.0% of the genome size. The presence of duplicate contigs was detected in the preliminary draft assembly of L. sibirica obtained in the second step, thus revealing the higher data ambiguity in the L. sibirica sequencing data compared to the A. thaliana data. To resolve the ambiguities in the second stage, the total number of all contigs resulting from the fifth set was increased by 16 folds by multiplying each contig 16 times, respectively. This trick allowed the CLC assembler to apply the majority rule when picking one of the alternative alleles, using the alleles selected in the fifths set in the first step of assembly. The same approach was used also for the Arabidopsis thaliana genome stepwise assembly by four different assemblers (Fig. 8 and Additional file 4: Table S4). The CLC Assembly Cell again demonstrated the best performance.Fig. 8

Results of the Arabidopsis thaliana genome stepwise assembly by four different assemblers using raw reads partitioned into five sets following the approach used for assembling of the Larix sibirica genome. Minimum contig length used for assembling was 200 bp

In addition, to verify the accuracy of the stepwise CLC Assembly Cell assembly the medium size genome (265 Mb, 2n =16) of Prunus persica (peach) was also assembled by both the traditional method using 24,324,216 sequence reads (~15X coverage) available on https://www.ncbi.nlm.nih.gov/bioproject/PRJNA31227 and the same stepwise approach that was used for the larch genome assembly and based on the five parts (Fig. 9 and Additional file 5: Table S5). The traditional and stepwise assemblies were similar to 95.64% based on the NUCMER comparison.Fig. 9

The traditional and stepwise CLC Assembly Cell genome assembly parameters for peach (Prunus persica). Minimum contig length used for assembling was 200 bp

From open-access databases, raw transcriptome data for 15 species were downloaded. Among the species, 12 species belong to 10 genera of Pinaceae, including Abies firma, Cathaya argyrophylla, Cedrus deodara, Keteleeria evelyniana, Larix gmelinii, Picea abies, Picea smithiana, Pinus armandii, Pinus elliottii, Pinus massoniana, Pinus taeda, Pseudolarix amabilis, Pseudotsuga menziesii, Tsuga dumosa and Tsuga longibracteata, and the three species Cycas panzhihuaensis, Araucaria cunninghamii, and Platycladus orientalis were used as outgroups (Supplementary Table S1).

We selected for our study the individual-based spatially explicit model LAVESI that was developed to simulate population dynamics of widespread Larix gmelinii in north-eastern Siberia (Kruse et al., 2016 ) and developed for treeline migration with a parametrisation for the Taymyr region (Kruse et al., 2016 ; Wieczorek et al., 2017 (link)). It simulates the life cycle of all trees in a given area for each year, starting from the seed stage. Hence, the model already includes the study species and was recently further equipped with new functions, which lay the groundwork for this study. Besides seed dispersal, wind-dependent and spatially explicit pollination was introduced (Kruse et al., 2018 ).
In short, the model simulates every year by going through the following sub-routines, established in previous publications of the model (Kruse et al., 2016 ; Kruse et al., 2018 ):
Initialisation: The environmental information (elevation, slope, terrain water index) of the simulated area is read in and filled in the model's structures. The weather provided in monthly temperature and precipitation values is read for a given number of years. Each simulation run starts with an empty area and seeds are introduced to the area during a spin-up period at the start of the model.
Environment is updated: As density map is calculated for the area in which the density influence is recorded for the trees. The competition between trees is calculated based on the basal diameter in a given area of the density map (Formula 1). The active-layer depth is estimated based on the number of days exceeding 0 °C. $densityinfluence\left(x-coord,y-coord\right)=\frac{diamete{r}_{basal}}{\sqrt{x-coor{d}^2+y-coor{d}^2}+1}$
Growth: The maximal growth is calculated based on an average of ten years’ climate data. The basal growth (Formula 2) of one individual for the year is then derived from this by including the density index of the tree. Based on the diameter, the tree height is estimated (Formula 3). $Growt{h}_{max,diameter}=\left(precipitation*Growt{h}_{standard,diameter}*\left(\frac{1}{f_{AATNDD}}*AATI+\left(1-\frac{1}{f_{AATNDD}}\right)*netdegreedays\right)\right)*\left(1-TDEI\right)$ $height=\{\begin{array}{cc}44.43163*diamete{r}_{basal}& ,height<1.3m\\ {\left(7.02*\sqrt{diamete{r}_{breast}}\right)}^2+130& ,height\ge 1.3m\end{array}$
Seed dispersal: Seeds that are still within cones are dispersed, the direction and distance are randomly determined influenced by wind data (Formula 4). When seeds leave the extent of the transect to either the east or west they are reintroduced from the opposite site, to simulate a larger forest and avoid the loss of many seeds. $\text{distance}=\sqrt{2{\left(\text{release}\text{height}·\frac{\text{wind}\text{speed}}{\text{fall}\text{speed}}\right)}^2\left(-\text{log}\left(\text{rand}\right)\right)+\frac{1}{2}\text{distanceratio}·\text{ran}{\mathrm{d}}^{-1.5}}$
Seed production: Once a tree has reached the height of maturation, based on a pre-generated distribution randomly assigned, it produces seeds. The amount produced in each year is based on the height of the tree, competition, and the weather (Formula 5). $\text{seeds}\text{produced}=\lfloor \text{facto}{\mathrm{r}}_S*\text{Diamete}{\mathrm{r}}_{\text{basal}}*\left(1.0-{\left(\frac{\text{height}}{50\mathrm{m}}\right)}^{-1.0}\right)\rfloor$
Establishment: Seeds that are on the ground germinate based on weather conditions (Formula 6). $probabilitytogerminate\left(year\right)=backgroundgerminationrate+\left(weatherqualityfactor*\frac{G_{rowthmax,basal}\left(year\right)}{Growt{h}_{standard,basal}}\right)$
Mortality: The probability of death is calculated for each tree and, based on that, it is semi-randomly determined whether the tree dies and is removed from the simulation. The calculation is based on long-term weather values, the calculated drought strength, competition of surrounding trees, the age and size of the tree, and a base mortality rate. For each of these a mortality value is calculated, these are then summed up and compared to a randomly generated number. If the sum of death probabilities is larger the tree dies. The death of seeds is determined at this step as well, although the mortality rate for these is fixed.
Ageing: The last step is an increase in the age of both the seeds and the trees. Since every year is simulated the age is advanced by once each cycle. The seeds are removed once they have reached a certain (3 years) age limit.

To predict post-fire stand structure and composition across a range of forests in the interior Pacific Northwest, we integrated datasets from remote sensing, field plots with pre- and post-fire measurements, contemporary forest inventories, and historical forest inventories and reconstructions in a novel modeling approach as shown in Fig 1. To incorporate and assess uncertainty of modeled estimates we used a Monte-Carlo simulation framework that carries estimates of uncertainty through modeling steps. First, we developed species-specific tree mortality models (hereafter “species-level models”) that relate probability of mortality to remotely sensed fire severity (RdNBR) and individual tree size from a field plot network in Oregon and Washington that experienced fire between measurement cycles. Next, we simulated fire by applying species-level models to individual trees within unburned stands in four National Forests in eastern Oregon to predict post-fire stand conditions across the observed range of RdNBR values (hereafter “stand-level models”). We then compared the results of simulated fire in contemporary stands to historical records and reconstructions of forest conditions to determine fire severity ranges that have the highest likelihood to restore historical conditions. Finally, we applied estimates from stand-level models to an example burned landscape to demonstrate the use of these methods for post-fire assessment and management (hereafter “landscape-scale model”).
We developed tree mortality models using a regional dataset and applied them to assess restorative fire severity ranges in eastern Oregon. The regional dataset primarily included plots from dry forest systems, but also included plots burned within fires in the Oregon Klamath Mountains, Oregon Cascades, and Washington North Cascades (Fig 1 in S1 Appendix). The eastern Oregon focal area is characterized by warm summers, cold winters, and precipitation falling mostly as snow [21 (link)]. Historically, fire-tolerant ponderosa pine was the dominant overstory species at lower elevations and co-occurred with western larch (Larix occidentalis) in northeastern Oregon. Less fire-tolerant white fir (Abies concolor), grand fir (Abies grandis), and Douglas-fir (Pseudotsuga menziesii) occurred as components of mixed stands at higher elevations and in more mesic topographic settings [67 (link)]. Before Euro-American colonization, frequent, low severity fire (8–31-year return intervals) maintained forests conditions that were relatively resistant to fire, drought, and native pathogens [21 (link), 68 (link)]. Fire exclusion due to decreased cultural burning, land use changes, and fire suppression policies has increased the abundance of less fire-tolerant species and overall forest density across these landscapes [21 (link), 65 (link)].