A Lysimeter Characterization of Evaporation by Eucalypt Forest and its Representativeness for the Local Environment

Abstract

A series of three 5–10 day comparative experiments in a eucalypt forest showed excellent agreement in hourly and daily evaporation rate between the forest around the site, as indicated by energy balance measurements, and a large monolith weighed lysimeter containing a representative sample of the local cover. Each time multiple concurrent neutron moisture water (NMM) readings, both within and external to the lysimeter, agreed well with the cumulative weight change and energy balance readings.

A further experiment took place some weeks after an emergency irrigation of the lysimeter, towards the end of a severe drought period. This revealed a major discrepancy between lysimeter and “external” evaporation, again supported by NMM readings, and explainable as due to a temporary anomaly in lysimeter soil moisture, with accompanying changes in plant growth notably in leaf area (LAI).

Such situations would normally be rare, and of comparatively short duration (in later experiments still, not reported here, full agreement was restored). Nevertheless the incident points to the need both to understand evaporation response to the major factors controlling it, as discussed in the paper, and to monitor these regularly as a basis for satisfactory extrapolation of lysimeter data.

Twelve months of near continuous lysimeter data permitted separate analysis of interception loss rates (E_I) and dry canopy losses (E). Lysimeter E_I was consistent with independent throughfall and stemflow measurements in the local forest, and on an event basis was found to approximate to a linear function of event rainfall.

With dry canopies, atmospheric demand or potential evaporation (Ep), soil water deficit (swd) and LAI were identified as major indices for E. Relating E to swd for various levels of Ep showed agreement with the model of Denmead and Shaw (1962), with progressive reduction in critical swd (the level at which actual evaporation drops sharply off from potential) as Ep increased. At LAI > 3, 86% of the variance in daily E was accounted for, using data for 102 dry sunny days. At lower LAI, the same behaviour was apparent but with lower levels of E.

In comparing eucalypt evaporation (E_euc) with that from coniferous forest and grassland (E_con and E_grass respectively), upper limits of E_euc on dry days were generally consistent with published values of E_con but less than E_grass under similar conditions - provided soil moisture was plentiful. However, under typical Australian conditions, with periodic shortages of soil water, seasonal or annual values of E_grass are likely to be less than for forests, due to shallow rooting.