Deciduousness in the New Zealand flora: Modelling the consequences of winter leaf loss for annual canopy carbon uptake

This thesis investigates aspects of deciduousness in the native New Zealand tree flora by modelling the consequences of winter leaf loss on annual carbon uptake for two co-occurring tree species with contrasting leaf habit; winter-deciduous fuchsia (Fuchsia excorticata) and annual-evergreen wineberry (Aristotelia serrata). Canopy carbon balance was modelled by incorporating measurements of cold-induced photoinhibition of photosynthesis, detailed descriptions of the phenology of leaf emergence and mortality, and the seasonal and temperature dependence of photosynthesis and respiration in a one-dimensional multi-layer model of canopy carbon uptake. Contrary to a widely held view, there was no evidence that variability of leaf loss in wineberry measured at a range of sites at the end of winter was related to site temperature. However, there were consistent patterns in leaf loss within wineberry canopies, with greater leaf-loss from open (66% of the total leaves emerged) than from shaded (54% of leaves emerged) shoots. It was hypothesised that this was a result of the adverse effects of high irradiance and low leaf temperatures during winter. Detailed measurements on a single tree supported this hypothesis. The efficiency of photosynthesis (measured via the in vivo ratio of variable to maximal chlorophyll a fluorescence, Fv/Fm) was lower in open than shaded leaves; the difference in Fv/Fm between leaves on open and shaded shoots ranged from 0.1 before the first frosts to 0.57 after leaf temperature of -4.7 DC. Fv/Fm declined markedly following frosts, to 20% of its pre-frost value in sun-exposed leaves, with Fv/Fm declines significantly (p < 0.001) related to high irradiance and low leaf temperature. Leaf loss increased following Fv/Fm declines, and was 0.22 and 0.07 leaves per shoot per day from open and shaded shoots, respectively. The phenology of leaf emergence and mortality for fuchsia and wineberry was quantified at a field site. Fuchsia leaf emergence started seven days sooner, was 0.06 leaves per shoot per day faster, and ended 40 days earlier than it did for wineberry. Similar patterns were observed for leaf mortality, with fuchsia experiencing the greatest rate of leaf loss (0.15 leaves per shoot per day) less than three months after budburst. Factors affecting fuchsia and wineberry leaf lifespans were investigated using failure-time analysis. Leaf survival probability decline to 50% 72 days after bud burst in fuchsia and 136 days after bud burst in wineberry. In comparison, 50% leaf mortality, estimated from sigmoid growth curves, had occurred by 107 and 202 days in fuchsia and wineberry respectively. The date of leaf emergence had a significant (p < 0.01) but opposite effect on leaf lifespan for both species; earlier emerged leaves had longer lifespans relative to later emerged leaves in fuchsia, whereas the reverse was true for wineberry. Leaf lifespans were significantly (p < 0.001) clustered at the shoot-scale, although this was not related to any shoot-level variables investigated (shoot height, branch order, number of leaves per shoot length). Responses of photosynthesis, and respiration at night, to CO₂ concentration and leaf temperature of wineberry and fuchsia leaves were measured seasonally at a field site to quantify parameters for a biochemical model of leaf photosynthesis. Parameters describing the maximum rates of rubisco carboxylation (Vcmax) and electron transport (Jmax) were significantly related to leaf nitrogen concentration (Narea), and were 60% higher on average in spring and summer leaves than they were in autumn and winter leaves for both species. These seasonal changes were not significantly related to changes in Narea. Values for Vcmax and Jmax were 30% lower in wineberry leaves than they were in fuchsia leaves on average, although this difference ranged from 15% in summer leaves to 39% in winter leaves. The measurements of photoinhibition, phenology, and photosynthesis and respiration were used as parameters in the annual carbon uptake model. For a wineberry-fuchsia forest dominated by wineberry, the model estimated net annual canopy carbon uptake (ANET) of 0.97 kg m-2, 93% of which was attributable to wineberry. Simulations where leaf emergence and mortality were advanced or delayed by up to 30 days decreased ANET by nearly 6% for fuchsia, whereas advancing emergence and mortality increased ANET by nearly 5% for wineberry. In simulations at a range of sites, an evergreen phenology always resulted in greater ANET than that estimated using the observed phenology in wineberry. A similar result was observed for fuchsia, except at sites where cold winter temperatures limited photosynthesis, where the reverse was true. While ANET can be greater with a deciduous phenology where winter cold limits uptake, the model simulations support the suggestion that the New Zealand flora has few deciduous tree species because mild winters mean that carbon uptake can continue year-round over much of the country. This shows that carbon balance considerations alone can account for the lack of deciduousness in the New Zealand flora