Applying a land systems approach to describe and partition soil and forest variablility, Southern Mamaku Plateau, part of Kinleith Forest, New Zealand

Abstract

Kinleith Forest is a P. radiata dominated plantation forest situated on the edge of the Taupo Volcanic Zone in the central North Island, New Zealand. Approximately 35 000 ha of the forest is situated on the southern Mamaku Plateau. The aim of this study was to determine the spatial structure of soil variability and integrate an analysis of the growth, productivity and nutrient variables for P. radiata to ascertain the detail of soil information appropriate for plantation forestry. A land systems approach, using generic and specific soil-landscape models and hierarchical landscape analysis, provides the framework for soil mapping. The approach is scale-flexible and lends itself to implementation of site specific forest soil, nutrient and health management practices.

The southern Mamaku Plateau consists of five superimposed Late Quaternary welded ignimbrites with the Mamaku Ignimbrite (c. 220 ± 10 ka) being the uppermost. Aeolian coverbeds consisting of predominantly loess and rhyolitic tephras mantle erosional surfaces formed in ignimbrite bedrock. Four episodes of strath cutting and valley incision into ignimbrite are recognised from the sequence of overlying loess and tephra coverbeds. The basal loess and/or tephra on the erosion surfaces denoting the cessation of each episode of strath cutting and valley incision are: (a) pre-Rotoehu loess, (b) pre-Kawakawa loess, (c) Rotorua Tephra, and (d) Taupo Ignimbrite.

The four episodes of strath cutting and valley incision can be synthesised into five topographically recognised stages of landform evolution. Landform evolution stage (I) is represented on the upper Mamaku terrain and stage (V) on the lower Mamaku terrain. In a revised coverbed distribution model the oldest coverbed sequence is recognised on hill and terrace land components on the upper and lower Mamaku terrain. The coverbed stratigraphy together with its soil stratigraphic interpretation and soil-landform relationships has been used to map nine land systems. Soil-landscape relationships and forest variability are described and analysed within the U3 complex land system. The U3 complex land system comprises Oruanui and Mamaku simple land systems, each with a common recurring pattern of hillock, ridge, flat and ravine land components, situated on the upper Mamaku terrain.

Soil and regolith variables were partitioned at progressively finer levels of hierarchical landscape subdivision to ascertain spatial variability structure. Taupo Ignimbrite thickness, topographically controlled over short range (10s of metres), and a rainfall (leaching) gradient over a longer range (1000s of metres) were identified as the main factors controlling soil variability. Land components within simple land systems effectively represented the variability. Simple and composite soils formed in Taupo Ignimbrite veneer deposits over yellowish-brown beds (collectively <2.5 m thick) dominate the U3 complex land system, with soils showing varying expression of podzolisation.

The structure of spatial variability of foliar nutrient concentration differs between foliar nutrients N, P, Mg, K and B. Complex land system for foliar N, P, Mg and K, and simple land system for foliar B were appropriate grainsize for amelioration purposes. Within U3, foliar P was deficient and foliar Mg was marginal with a high probability of deficiency. Forest productivity was measured using site index, basal area index and volume index. Basal area index was the most effective indicator of forest productivity variability. Spatial variability was best partitioned by land components. Spatial variability of foliar nutrients and forest productivity indices were not structured in the same way, indicating a weak correlation between foliar nutrient variables and forest productivity.

The relationships between P. radiata and soil chemical properties were established by a pot trial for soils on the flats within U3. Principal Component Analysis (PCA) identified the main effects on P. radiata shoot growth and nutrient uptake as nutrient availability, P sorption and Mg inhibition. Podzolisation, although evident in the soil morphology, was not shown to significantly influence shoot variables. Shoot dry weight and height were significantly greater for seedlings grown in A horizons compared to seedlings grown in the Bw and BC horizons in Taupo Ignimbrite and buried Bw horizons in yellowish-brown beds. The nutrient availability effect was most apparent in A horizons, while the Mg inhibiting effect and P sorption effect were responsible for limiting shoot growth in the Taupo Ignimbrite and yellowish-brown beds respectively.

A nutrient availability index (NAI), developed from the pot trial and soil horizon stratigraphy, was calculated for hillock and flat land components in the Oruanui and Mamaku land systems. Poor correlation between NAI and site productivity (basal area index) reaffirmed the influence of non-soil related factors on forest productivity variability.

The land systems approach effectively partitioned soil and forest spatial variability at different levels of resolution. However, soil and soil-induced forest variability are not strong because of the homogenising effects of widespread aeolian (including pyroclastic) coverbeds.

Implementing a scale-flexible mapping approach, such as land systems, would provide a basis for improved inventory data collection, and management of forest soils, nutrients and health in Kinleith Forest.