Leo Goudzwaard

Controls on Coarse Wood Decay in Temperate Tree Species: Birth of the LOGLIFE Experiment

Dead wood provides a huge terrestrial carbon stock and a habitat to wide-ranging organisms during its decay. Our brief review highlights that, in order to understand environmental change impacts on these functions, we need to quantify the contributions of different interacting biotic and abiotic drivers to wood decomposition. LOGLIFE is a new long-term ‘common-garden’ experiment to disentangle the effects of species’ wood traits and site-related environmental drivers on wood decomposition dynamics and its associated diversity of microbial and invertebrate communities. This experiment is firmly rooted in pioneering experiments under the directorship of Terry Callaghan at Abisko Research Station, Sweden. LOGLIFE features two contrasting forest sites in the Netherlands, each hosting a similar set of coarse logs and branches of 10 tree species. LOGLIFE welcomes other researchers to test further questions concerning coarse wood decay that will also help to optimise forest management in view of carbon sequestration and biodiversity conservation.

Flood-Ring Formation and Root Development in Response to Experimental Flooding of Young Quercus robur Trees

Spring flooding in riparian forests can cause significant reductions in earlywood-vessel size in submerged stem parts of ring-porous tree species, leading to the presence of ‘flood rings’ that can be used as a proxy to reconstruct past flooding events, potentially over millennia. The mechanism of flood-ring formation and the relation with timing and duration of flooding are still to be elucidated. In this study, we experimentally flooded 4-year-old Quercus robur trees at three spring phenophases (late bud dormancy, budswell, and internode expansion) and over different flooding durations (2, 4, and 6 weeks) to a stem height of 50 cm. The effect of flooding on root and vessel development was assessed immediately after the flooding treatment and at the end of the growing season. Ring width and earlywood-vessel size and density were measured at 25- and 75-cm stem height and collapsed vessels were recorded. Stem flooding inhibited earlywood-vessel development in flooded stem parts. In addition, flooding upon budswell and internode expansion led to collapsed earlywood vessels below the water level. At the end of the growing season, mean earlywood-vessel size in the flooded stem parts (upon budswell and internode expansion) was always reduced by approximately 50% compared to non-flooded stem parts and 55% compared to control trees. This reduction was already present 2 weeks after flooding and occurred independent of flooding duration. Stem and root flooding were associated with significant root dieback after 4 and 6 weeks and mean radial growth was always reduced with increasing flooding duration. By comparing stem and root flooding, we conclude that flood rings only occur after stem flooding. As earlywood-vessel development was hampered during flooding, a considerable number of narrow earlywood vessels present later in the season, must have been formed after the actual flooding events. Our study indicates that root dieback, together with strongly reduced hydraulic conductivity due to anomalously narrow earlywood vessels in flooded stem parts, contribute to reduced radial growth after flooding events. Our findings support the value of flood rings to reconstruct spring flooding events that occurred prior to instrumental flood records.

Faunal community consequence of interspecific bark trait dissimilarity in early-stage decomposing logs

Dead tree trunks have significant ecosystem functions related to biodiversity and biogeochemical cycles. When lying on the soil surface, they are colonized by an array of invertebrate fauna, but what determines their community composition is still unclear. We apply community assembly theory to colonization of tree logs by invertebrates. During early decomposition, the attached bark is critically important as an environment filter for community assembly through habitat provision. Specifically, we hypothesized that the more dissimilar bark traits were between tree species, the more their faunal community compositions would differ. We tested this hypothesis by investigating the effects of bark traits on the invertebrate communities in the early-decomposing logs of 11 common, temperate tree species placed in the ‘common garden’ experiment LOGLIFE. Bark traits included bark looseness, fissure index, outer bark thickness, ratio of inner to outer bark thickness, punch resistance, water storage capacity and bark pH. The predominant faunal groups studied were Annelida, Isopoda, Chilopoda, Diplopoda, Diptera and Coleoptera. Our results showed (i) strong interspecific differences in bark traits, (ii) that bark traits related to environmental buffering had profound effects on the abundance of specific invertebrate groups, and (iii) the higher the overall bark trait dissimilarity between tree species, the more dissimilar these tree species were in faunal community composition, and the higher was the joint invertebrate family richness. A suite of bark traits together has fundamental afterlife effects on invertebrate community assembly, strongly filtering the colonizing invertebrates in early-decomposing logs, driving variation in their community composition and diversity. Our findings indicate that bark trait dissimilarity among tree species in forest stands is likely a better indicator of early-phase dead trunk fauna diversity than tree species diversity per se. A lay summary is available for this article.

Groei en productie van grove den in Nederland

In the Netherlands, growth and yield research on Scots pine was done from 1949 to 2002. This includes studies by Becking and by the Dorschkamp/IBN research institute. Together with the permanent sample plots from the timber prognosis system HOSP, all this comprises a dataset of 94 plots with 348 recordings. For the development of top height htop with age (t), Cieszewski’s model with site index h70 and 3 additional parameters fitted best. The diameter development up to stand height of 7 m was best described with the model by Jansen et al. based on htop and initial density N0. From a stand height of 7 m and up, the basal area increment iG was best described by a power func-tion based on htop, calendar year (yor), h50 and the stand density index of Hart (S%). For S % > 20.6 the basal area increment drops strongly with increasing S %. The effect of thinning on diameter after thinning was modelled with a modified La Bastide-Faber model. With all models together, a stand projection model was constructed, which follows the measured stand development reasonably well. The model was used to construct yield tables with five site classes and four thinning intensities.

Groei en productie van beuk in Nederland

In the Netherlands growth and yield research on common beech was done from 1960 to 1992 at very small scale. This includes studies by Becking and by the Dorschkamp/IBN research insti-tute. Together with the permanent sample plots from the timber prognosis system HOSP, all this comprises a dataset of 47 plots with 169 recordings. For the development of top height (htop) with age (t) Cieszewski’s model with site index h70 and three additional parameters fitted best. The diame-ter development up to stand height of 7 m was described with the model by Jansen et al. based on htop and initial density (N0). From a stand height of 7 m and up, the basal area increment (iG) was also described by a model from Jansen et al., based on a power function with htop, t, year of recording (yor), and the stand density of Hart (S %). For S % > 17.6 the basal area increment drops strongly with increasing S %. The model contains a correction factor for yor, which unfortunately could not be esti-mated from the plot data. The effect of thinning on the diameter after thinning was modelled with a modified La Bastide-Faber model. With all models together, a stand projection model was con-structed, which describes the measured stand development moderately well. The model was used to construct yield tables for with five site classes and six thinning intensities

Groei en productie van es in Nederland

In the Netherlands growth and yield research on ash was done from 1949 to 1988. This in-cludes studies by the Dorschkamp/IBN research institute and by Wageningen University. Together with the permanent sample plots from the timber prognosis system HOSP, all this comprises a da-taset of 41 plots with 150 recordings. For the development of top height htop with age t, Cieszewski’s polymorphic model with site index h50 and three additional parameters fitted best.The diameter development up to stand height of 7 m was described with a linear function in htop and tree distance. From a stand height of 7 m and up, the basal area increment (iG) was described by Jan-sen’s et al. model based on a power function with h50, htop, t, year of recording (yor), and the stand density of Hart (S %). For S% > 16.7 the basal area increment strongly decreases non-linearly with in-creasing %. The model contains a correction factor for yor, but this was not significant, nor was htop. The effect of thinning on the diameter after thinning was modelled with a modified La Bastide-Faber model. The models were used to construct yield tables for with five site classes and one thinning in-tensity.

Groei en productie van populier in Nederland

Between 1947 and 2000, growth and yield of Poplar was studied in the Netherlands. To the permanent plots measured by Becking and De Dorschkamp/IBN, the permanent sample plots from HOSP were added, resulting in 235 plots and 1808 recordings.The development of the mean height hm with age t was found to be best described by a Chapman-Richards-model, with adjustment factors for dense stand and for different parameters for forest stands, line plantings, and for aspen, using site index h25. The diameter development up to a height of 7 m was best described with a Gompertz-model in t, hm and the initial spacing (SP0).The basal area increment IG was best explained by a power function in which hm, age, h25 and the Becking-Hart spacing index S% are included. For S% > 29.2 the basal area increment dropped with a non-linear function in S%. For dense spacing and for line plantings, other models fitted best.Combining all models, a stand projection model was constructed, which described the plot develop-ment as measured reasonably well.Yield tables were made for forests with different spacing, for forests with a wide stand with and with-out systematically thinning, for forest with very dense spacing including mortality, for line plantings without thinning, and for aspen with heavy thinning from below