Regeneration and early tending of black locust (Robinia pseudoacacia L.) stands in the north-west of Romania

Aim: The aim of this study is to highlight the importance of black locust (Robinia pseudoacacia L.) on extreme site condition like sand dunes. Aspects of natural regeneration and early results of tending operation in terms of quality, growth and yield of young black locust stand are presented. Methodology: A Research and Development project was carried out in Carei-Valea lui Mihai Plain (north-west of the country) since 2016. Three management units were selected: pure natural regeneration by root suckers of black locust at different ages (management units 3B and 52A%) and a mixed black locust-black cherry stand (management unit 23D). Biometrical measurements and analyses as well as biomass estimations were performed. Results: (1) Black locust was established naturally by root suckers and the stocking of newly established stands can be as high as 50,000 suckers ha; (2) The initial growth of black locust regeneration is quick and the young regeneration can close the canopy in 1-2 years, resulting in an appropriate dune fixation and wind erosion control; (3) The young pure or mixed black locust-dominated stands are left untended until the first cleaning-respacing (mean diameter 5-6 cm), when the stand shows high stocking/density and a wide variation in tree size. This intervention is from below, heavy (intensity over 25% by number of trees or basal area) and of negative selection type, removing mostly low Kraft class, dead or dying, and defective trees. Interpretation: The potential of black locust to establish naturally by root suckers and their fast initial growth, despite the unfavourable conditions in the case study area, resulted in effective dune stabilization and wind erosion control. The effect of cleaning-respacing on the early growth of this species in pure or mixed stands was outlined.


Black locust in the world, in Europe and in Romania
Black locust (Robinia pseudoacacia L.) originates from the eastern half of the United States, where it is found in two areas, eastern (Pennsylvania, Ohio, Alabama, Georgia and South Carolina states) and western (Missouri, Arkansas, and Oklahoma states) (Huntley, 1990).
The area covered by black locust in 1922 was only 28,000 ha (Drăcea, 1928), expanding to ca. 100,000 ha by the mid-1950s (Negulescu and Săvulescu, 1957) and further to approx. 250,000 ha at the present time. 4% of this is on national forest land, mostly in the south of the country, on sand dunes and areas with heavy soils in the forest steppe zone (Enescu and Dănescu, 2013;Nicolescu et al., 2018).

Regeneration and early growth of black locust
In different parts of the world, black locust is regenerated by one of three methods: a. Planting in spring using 1-year-old seedlings, normally bare-rooted, 0.5-1.0 (or even 2) m tall, produced in conventional nurseries (Olson, 1974;Luna, 1996;Redei, 2013a). The initial stocking rate of black locust plantations in Europe is very variable: 1,100-1,900 seedlings ha -1 in France (Bourgogne) (Borde, 2011), 1,200-1,700 seedlings ha -1 (4 m x 2 m, or 3 m x 2 m) in France (Aquitaine and Poitou-Charentes) (Carbonnière et al., 2007;CRPF, 2007), 2,000-2,500 seedlings ha -1 (2.5 x 2.0 m, or 2.5 x 1.6 m) in Poland (Wojda et al., 2015) to 4,000-6 5,000 seedlings ha -1 (2.0 m x 1.25 m, or 2.0 x 1.0 m in Romania; 2.4 m x 0.7-0.8 m, or 2.4 m x 1.0 m in Hungary) (MAPPM, 2000a;Rédei et al., 2008;Redei et al., 2011;Rédei, 2013a). b. Naturally by seed. This is rare, as the hard and impermeable seed coat limits germination in the forest/natural environment. However, there are some examples of natural regeneration in the literature (Negulescu and Săvulescu, 1957;Stănescu, 1979), this process being facilitated by seed wounding with heavy machinery, or natural thermal shock (Stănescu et al., 1997;Șofletea and Curtu, 2007). c. Naturally by vegetative regeneration from stool shoots and root suckers. As black locust coppices freely this is considered the most cost-effective management system for the species (Bîrlănescu and Belu, 1968;Sitzia et al., 2016). The method is cheap, efficient and allows local people to collect the stem wood, highly valued as firewood. Root suckers live longer and are healthier (i.e. show less rot at the same age) than stool shoots, however the latter grow quicker up to 12-15 years of age than root suckers (Negulescu and Săvulescu, 1957;Haralamb, 1967;Stănescu, 1979). The most common rejuvenation method is by root suckers as black locust develops horizontal, shallow and wide-spreading roots which can extend 15-20 m from the parent tree (Negulescu and Săvulescu, 1957;Stănescu, 1979;Luna, 1996).

Early management of black locust stands
The application of early management operations, such as release cutting and cleaningrespacing, in black locust stands varies according to the regeneration method as follows: No release cutting is needed in plantations but it is necessary in coppice stands to reduce the number of shoots per stool to 1 or 2 and to protect root suckers from stool shoot competition (Haralamb, 1967;Debenne, 1988).
a. In plantations with up to 5,000 seedlings ha -1 there is no need for any release cutting (MAPPM, 2000b;CRPF, 2007;Wojda et al., 2015). In such stands cleaning-respacing begins after canopy closure, at 4-5 years and the stocking should be reduced to about 2,500 trees ha -1 (Wojda et al., 2015). The second cleaning follows 2-3 years later, with a further reduction to ca 1,700 trees ha -1 (Wojda et al., 2015).
b. In black locust coppice stands regenerated from stool shoots and root suckers, release cutting is necessary to reduce the number of shoots per stool to 1 or 2 and to protect root suckers from stool shoot competition (Haralamb, 1967;Debenne, 1988). Normally two release cuttings are performed, the first one in the first or second year, followed by another 1-3 years subsequently (Costea et al., 1969;MAPPM, 2000b). In Romania, two cleaningrespacing operations are performed in years 3-4 and 6-7, reducing the canopy cover to 80-85%.
In both black locust plantations and coppice stands cleaning-respacing is considered "the basis for all good management in black locust stands" (Halupa and Rédei, 1988). These authors aimed to heavily reduce the number of stems, allowing the potential final crop trees sufficient space to grow. If this intervention is too late or too light, the remaining trees were found not to develop their crowns normally (they are deformed or very small) as this is a strong light demanding species and is intolerant of shade/competition (Negulescu and Săvulescu, 1957;McAlister,1971;Harlow et al., 1986). They do not recover no matter how much light is subsequently supplied (Haralamb, 1967;MAPPM, 2000b). The cleaningrespacing is based on negative selections (particularly in the first intervention) removing defective trees, for example those that are forked (this species is sensitive to early frosts, leading to forking - Haralamb, 1967;Constantinescu, 1976), badly formed, wounded, bentover (the effect of strong phototropism) combined with positive selections where even wellformed and healthy individuals are removed to provide additional growing space to those selected to remain (Constantinescu, 1976). Halupa and Rédei (1988) highlighted the importance of cleaning-respacing to produce regular spacing of the remaining trees.
In the context of these characteristics of black locust stands and silviculture in the early stages, a Research and Development (R&D) project was launched in 2016, in order to evaluate the regeneration and early tending of black locust stands in the north-west of Romania. The objectives of this project are (1) to assess the regeneration potential of black 8 locust by root suckers, (2) to assess the early growth of root sucker stems and (3) to follow and evaluate the early results of these interventions in terms of quality, growth and yield of young black locust stands.
According to Spîrchez et al. (1962) and Târziu and Spârchez (2013), the local soils are part of the Psamments suborder, Entisols order (sandy soils), with the following characteristics: (i) very deep but poor, with a low fertility and low nutrient (N, P, and K) content; (ii) light soil texture (85-90% sand, mostly fine); (iii) moderately acid to neutral (5 to 7) pH; (iv) maximum fraction of humus is 1% in the upper 25 cm of soil; (v) presence of a hard and poorly drained ortstein (ironpan) horizon, Al, Fe, Mn, and humus compounds-rich from the overlying shallow O horizon. The ironpan is sand-cemented and developed on the dune top and along the dune slopes from 55 cm deep downwards (Spîrchez et al., 1962). This horizon restricts water infiltration during the driest summer periods, when the sand gets very warm at the surface and to a considerable depth, providing an important water supply for the forest vegetation.
The local climate is classified as temperate-continental (C.f.b.x. type according to Köppen -ICAS 2008), compared to a humid climate in the native range of black locust (Huntley, 1990).
Mean annual temperature: 10.3 C o ; maximum monthly temperatures in July (20.7 C o ), minimum in January (-1.6 C o ). Mean annual total precipitation: 573.3 mm, and this value is much lower than the one found in the U.S.A., ranging between 1020 and 1830 mm (Huntley, 1990). The maximum monthly precipitations is in June (83 mm (Rițiu et al., 1988). Mean length of frost-free period is 270 days, much longer than in the native range, where it is between 150 and-210 days (Huntley, 1990). The mean annual aridity (de Martonne) index is 28.2, so the area is considered to be located in the transition zone between the plain forest zone, moderately humid, and the forest steppe zone.
Until 1933 only small-scale plantations including Scots pine (Pinus sylvestris), black pine (Pinus nigra), pedunculate oak (Quercus robur), northern red oak (Q. rubra), pin oak (Q.  1960-1980 (450 ha), with the majority on low fertility former agricultural land (Spîrchez et al., 1962). Currently forest vegetation covers about 12% of the total area of the Carei-Valea lui Mihai Plain, with black locust the main tree species covering over 80% of the total forest area (ca 3,000 ha). Other tree species, found on a much smaller scale in the same area, are black cherry, Canadian poplars (Populus x euramericana), oaks (e.g., pedunculate, northern red, pin), and pines (black, Scots).
In the Carei-Valea lui Mihai Plain, black locust has been used since 1892 on (i) low-fertility former agricultural land or (ii) for replacing low-productive tree species such as Quercus robur, Tilia cordata, Acer campestre, and Ulmus campestris, in order to prevent wind erosion and sand dune movement and to produce firewood. Subsequently, black locust stands were treated as simple (low) coppice, usually on a rotation of 20-30, but up to 35 years, similar to the timescales found in the U.S.A. (McAlister, 1971), India (Luna, 1996), France (Carbonnière et al., 2007;Borde, 2011) and Hungary (Rédei et al., 2011;Rédei, 2013c), depending on yield class.

Experimental material
In this context, three management units -3B, 23D and 52A% -were selected for the R&D project. The main characteristics of these stands are: (1) Management unit 3B -pure natural regeneration by root suckers of black locust, 1-year-old, following simple coppice cut (winter 2015-2016) and removal of stumps; (2) Management unit 52A% -pure natural regeneration by root suckers of black locust, 2-years old, following simple coppice cut (winter 2013-2014) and removal of stumps; (3) Management unit 23D -mixed black locust-black cherry stand, 12-year-old, originating from root suckers after a simple coppice cut (2004) and removal of stumps. No silvicultural interventions had been performed since establishment.

Root suckers/Trees measurements
Root collar diameter and total height were measured for all initial and remaining root suckers, after the release cutting carried out in all plots from management units 3B and 52A%.
Diameter at breast height (DBH) and total height for all initial and remaining trees after the cleaning-respacing were measured in management unit 23D. The location (x-y) of each remaining tree as well as four perpendicular crown radii for such trees were also measured.

Biomass estimation
The suckers cut during the release intervention (management unit 3B) were bundled in each plot and transported to the laboratory. The dry matter content was determined by drying material at 105 0 C, until constant weight. To assess biomass production, an allometric relationship between W and stem diameter was used: = , where W -biomass, Droot collar diameter, b and cconstant parameters.

Silvicultural, biometrical and statistical analysis
Using the data collected in the field the following calculations were performed: stocking (no. of trees/ha) before and after release cutting and cleaning-respacing, in order to determine the intensity of the interventions; density (sq.m ha -1 ) before and after cleaning-respacing, with the same purpose; mean collar diameter in release cutting, DBH in cleaning-respacing, of initial, extracted and remaining trees and their standard deviations; mean height (in both release cutting and cleaning-respacing) of initial, extracted and remaining trees and their standard deviations; dry biomass of extracted, remaining and initial suckers in management unit 3B; coefficients of variation of diameters and heights (initial, extracted and remaining suckers or trees); significant differences beetwen means were tested using ANOVA and Duncan post hoc test (Statistica 7, Statsoft Inc, USA); differences between means were analysed using ANOVA and the significant effects were tested with Duncan post hoc test.

Results and discussions (A) Regeneration of black locust stands
In the two management units analyzed with respect to the natural regeneration of black locust by root suckers, the most relevant results are as follows:

(i) 1 year old natural regeneration (management unit 3B)
The potential for natural regeneration of black locust from root suckers was very high and the initial stocking after one growing season ranged between 15,200 and 67,600 suckers ha -1 (50,800 on average). The initial stocking was higher than that found in the study made in France, which recorded over 40,000 suckers ha -1 using this regeneration method (Pagès, 1985, cited in Carbonnière et al., 2007. This very high stocking allowed for very heavy interventions (over 25% of the number of trees -MAPPM, 2000b) with release cutting (from 68.42% to 91.89%, over 80% in the majority of plots), reducing the stocking per ha to between 4,800 suckers and 9,200 suckers (7,200 suckers on average) (Table 1). The suckers removed by release cutting were the smallest (or thinnest) ones. Consequently, the arithmetic mean collar diameter of black locust suckers increased from 7.36±3.91-9.81±5.32 mm to 11.67±4.12-14.72±6.76 mm, the coefficients of variation of diameters being the lowest in the remaining root suckers. No significant differences were registrated for the remaining root suckers (F=1.010, p=0.416) ( Table 2). The suckers extracted through this intervention were also the shortest, so the arithmetic mean height of black locust suckers increased from 87.79±40.18-100.18±63.29 cm to between 126.33±35.06 and 175.67±44.88 cm, the coefficients of variation of heights being also the smallest in the remaining root suckers. In terms of root sucker height, the analysis of variance showed no significant differences: F=0.896, p=0.483 for initial root suckers, F=1.193, p=0.311 for removed root suckers and F=1.469, p=0.207 for the remaining ones.
The aboveground dry biomass of initial, extracted, and remaining black locust root suckers in the six plots was calculated using the allometric formula = 0.0652 2.582 , R 2 =0.9426.
The initial above ground dry biomass in all plots, except plot no. 4, exceeded 1.1 t ha -1 (1.384 t ha -1 on average), the maximum being measured in plots 3 (1.875 t ha -1 ) and 6 (1.939 t ha -1 ) (Table 3), which were both located close to the dune top. As the release cutting intervention had a very heavy intensity, the remaining above ground biomass was less than 0.5 t ha -1 , with the exception of plots 3 (0.699 t ha -1 ) and 6 (0.565 t ha -1 ), with a mean of 0.462 t ha -1 .

(ii) 2 years old natural regeneration (management unit 52A%)
This stand was regenerated identically to management unit 3B and is located in very similar ecological conditions. In the two plots the very strong competition between the suckers started immediately after the canopy closure of the newly established regeneration, i.e. at the end of first growing season, producing an abrupt reduction in stocking of this 2-year-old stand (12,000 suckers ha -1 in plot 1 and 22,000 suckers ha -1 in plot 2) ( Table 4). The mean collar diameters (6.9 mm and 6.7 mm respectively) and mean height (162 cm and 155 cm respectively) are similar in the two plots. The ranges in both these parameters are similar and no significant statistical differences occurred (F=0.132, p=0.717 for collar diameter and F=0.372, p=0.543 for height) between the collar diameter and height of suckers in these two plots (Table 4).

(B) Cleaning-respacing of young mixed black locust-dominated stands
This intervention was carried out in management unit 23D, which showed the following main characteristics: The stand initial stocking was very high (5,467 trees ha -1 ) in plot 1, and lower (3,533 trees ha -1 ), in plot 2. These trees had a similar basal area (14.30 sq.m ha -1 in plot 1 and 13.87 sq.m ha -1 in plot 2). This made possible a very heavy intervention, with removal of over 25%,

17
reducing the stocking to 2,333 trees ha -1 and the basal area to 9.33 sq.m ha -1 (plot 1) and 1,733 trees ha -1 and 9.10 sq.m ha -1 (plot 2) ( Table 5). As the intensity by number of trees (57.3% in plot 1, and 50.9% in plot 2) was much higher than by basal area (34.1% in plot 1 and 34.4 in plot 2), the intervention was from below in both plots, removing mostly trees from the lower diameter classes.
As the intervention removed mostly the smallest (thinnest and shortest) trees, the arithmetic mean diameter and arithmetic mean height increased in both black locust and black cherry, particularly in the former species (Table 6). The intervention produced gaps in the canopy cover, which shows a value after cleaningrespacing of circa. 80% in plot 1, and 75% in plot 2, so that some trees have additional space at canopy level to develop their crowns and consequently increase DBH.

Conclusions
The R&D project, which began in 2016, focusing on pure and mixed black locust-dominated stands has led the following conclusions on regeneration and the early management of such stands: • The potential of black locust to establish naturally by root suckers after a low coppice cut and stump removal is very high and the stocking of such newly established stands can exceed 50,000 suckers ha -1 .
• Despite the unfavourable conditions in the case study area, the initial growth of regenerated black locust is fast and the newly established stand can close the canopy in 1-2 years, resulting in effective dune stabilization and wind erosion control.
• There are significant biometric differences, for example in collar diameter and height, between the young shoots, leading to a high level of natural mortality after canopy closure.
• Economic factors, such as lack of markets and/or workforce, results in young pure or mixed black locust-dominated stands usually being untended in the early stages. The first commercial intervention (cleaning-respacing) occurs when the stand has reached the thicket stage (minimum mean diameter 5-6 cm) and exhibits high stocking and density as well as wide dimensional (DBH and ht) variation. Consequently, the first cleaning-respacing intervention is from below, of high intensity and negative selection type, removing mostly low Kraft class (intermediate/suppressed), dead or dying, and defective (for example forked, wounded, or bent-over ) trees.
However, these are only preliminary results and at the next intervention, different measurements (e.g. collar diameters and heights -stands for release cutting; diameters, heights and crown radii -stands for cleaning-respacing) will be taken. These will provide an assessment of natural dieback in young, naturally regenerated, black locust stands, and the effects of the two silvicultural interventions on the early growth of this species in pure or mixed stands.