Modelling of root growth of corn (Zea mays L.) under fiield conditions

Abstract

Quantitative studies of root development as it depends on climate, soil and plant properties are one prerequisite to derive relations that allow for a more general, possibly predictive approach to root growth and, consequently, root nutrient uptake. Since net maize root length, e.g. amount of roots that can be measured or observed in a soil sample, increases until flowering, detailed information is needed until this stage is entered. The process of dispersion of the total amount of roots formed in soil, i.e. root penetration into soil and root proliferation in soil layers, has to be quantified.

Five field experiments, conducted in three years substantially differing in climate are used to describe root and shoot development on a silt loam soil at one location (West Lafayette, Indiana, USA). Both, total root length development and shoot growth of maize were related to the product of accumulated thermal time and solar radiation for all 5 experiments. Root growth in the topsoil layer was positively related to soil water content as calculated with a mechanistic soil water balance model.

Regression analysis of a long term experiment showed that on a silt loam soil there was a positive relation between root length at flowering in the topsoil layer and the amount of rainfall during the last weeks prior to flowering, e.g. the time of fastest root growth. Root length accumulation at the bottom of the soil profile was positively related to air temperature during the weeks following planting.

Experiments conducted in thin-layer cuvettes filled with a loamy sand soil allowed for a non-destructive and repeated observation of plant root development. The development of a software program enabled quantification of root development on the transparent front of the cuvettes. After penetration of a soil layer root growth was exponential followed by a linear growth phase. Secondary roots were formed at seminal roots approx. 2 days old, and secondary root elongation continued for 6 to 8 days. Only few tertiary roots were observed on scanner images. With decreasing depth the maximum absolute growth rate of roots decreased. Water content of soil layers (SWC) influenced penetration rate of roots into soil as well root proliferation within soil layers. Decreasing SWC increased number of secondary roots and reduced length of secondary roots, the overall effect being a reduction in absolute root growth rate. Increased soil bulk density (SBD) reduced total root length produced in cuvettes. The main reason was reduction of number of secondary roots formed. Seminal root development was not strongly influenced by soil bulk density. With decreasing depth the effect of SBD on the formation of lateral roots increased strongly.

It has been shown that during vegetative growth of the corn plant shoot growth and root abundance are interdependent, and a maximum root-shoot ratio (RSR) can be derived. Root penetration into soil and root proliferation during the exponential and linear growth phase in soil layers (root dispersion) can be described sufficiently exact by regression equations that take into account air temperature and solar radiation. A maximum absolute penetration rate (APR) and growth rate (AGR) can be derived. AGR and APR are influenced by distance from soil surface, soil water content and soil bulk density; these relations can be expressed mathematically.

On basis of the results obtained a one-dimensional mathematical model is developed for a layered soil using STELLA software. Input data are daily values for minimum and maximum air temperature, and solar radiation. The soil profile is subdivided in up to 10 layers with a minimum depth of 5 cm, and each layer is characterized by one soil bulk density. Daily values of soil water contents of soil layers are calculated using an external model of soil water balance. From these data the model calculates shoot biomass accumulation, root abundance from shoot growth, root penetration into soil, and root proliferation in soil layers. Maximum absolute root growth rate and maximum penetration rate are altered with soil depth, with changes of soil water content for soil layers over time, and as soil bulk density differs for soil layers.

The following rules were implemented:

• 1. Root length produced per day cannot exceed the value resulting from shoot growth and RSR, e.g. maximum root growth is strictly dependent upon shoot biomass accumulation. Weather conditions alter shoot growth and hence alter root abundance.

• 2. Root proliferation in any one-soil layer cannot start before the root has penetrated to that depth of the soil profile. The sum of root length produced in all soil layers with roots cannot exceed maximum root abundance.

• 3. In each soil layer root growth starts with an exponential phase followed by a linear phase later on. One maximum AGR of the root applies to all soil layers.

As could be expected for the data set used for model development, calculations agreed well with data of shoot growth and root abundance. Root growth in soil layers was best reflected for the 0–15 and 15–30 cm layer, and deviations between measured and predicted values were highest for the deepest soil layers.

It is concluded that the root growth model reflects the major processes that deter­mine root abundance and dispersion for the weather and soil conditions that are characteristic for the location. Inclusion of other soil types and climates may extend the applicability of the model to other environmental conditions in the future.

The model presented here is simplistic in a number of aspects. The following list gives some keywords and is not final:

• The root growth model is one-dimensional and roots are evenly distributed in the soil volume; clustering of roots and possible effects on nutrient and water acquisition are not represented.

• The soil related algorithms incorporated in the root model are site specific; the influence of soil types differing in pore space, pore diameter and pore continuity are not represented.

• The possible influence of adverse soil chemical conditions (local response) is not incorporated; salinity effects and aluminum toxicity are not represented.

• Turnover processes in soil that are mediated by microorganisms are not included; root decay, nitrogen liberalization are not represented.

• Processes that influence nutrient availability other than physico-chemical are not included, e.g. effects of root exudation and mycorrhiza infection are not represented.

• Different binding forms of ions in soil are not implemented, e.g. non-exchangeable potassium that has been shown to be available to plants, is not represented.

• The influence of heat, cold and drought stresses on shoot (and root) growth are not represented.

The list shows that model calculations cannot be expected to agree well with observations if one or more of these factors are dominating over the processes that are at present incorporated into the model. However, the degree of agreement may be used to decide whether or not additional processes have to be included. The presented, simplistic model can be used as a starting point.