REDUCING SURFACE ROOTING OF TREES WITH CONTROL PLANTERS AND WELLS

Control of surface rooting was explored in a study of fruitless mulberry (Morus alba) and zelkova (Zelkova serrata). After 3'A growing seasons, average amounts of roots were reduced substantially in the top 8 inches of soil by PVC control planters and by wells that placed trees 18 inches below grade. Among trees in planters and wells, however, amounts of surface roots differed greatly, and the tendency of roots to return to surface layers may be partly associated with soil compaction and poor aeration. Mulberry roots returning to the surface layers were, on the average, smaller and zelkova roots fewer than the unconstrained roots of control trees of the two species. For a given trunk size, mulberry roots were much more massive than zelkova roots. Avoiding species with massive roots remains extremely important in selecting street trees. One of the costliest operations associated with municipal trees is repairing sidewalks damaged by tree roots. The cost of repairing damage by one tree is often $500 or more per occasion. And, once started, damage often recurs at about 5-year intervals. If cities do not repair the walks, they may face even higher costs when sued for injuries caused by broken or misaligned sidewalks (4, 8). Various efforts have been made to control root damage to sidewalks. Some cities plant trees below grade in wells so roots at least begin at greater depths. Rigid PVC plastic control planters and other barriers to contain roots have been developed, and, in Australia, polyethylene sheet plastic and fumigants have been tested for stopping roots (6). This paper reports a study of wells and control planters for controlling surface roots of fruitless mulberry and zelkova trees. Sidewalks and soil amendment effects on root development were also examined. Procedures In 1980, in cooperation with the University of California Department of Environmental Horticulture, we planted 32 fruitless mulberry trees (Morus alba) and 32 zelkova trees (Zelkova serrata) in a formal experimental arrangement, along with 8 extra trees for practice excavation and possible replacement of trees that might die. Trees were arranged in 12 rows of 6 trees each with 18-foot spacing between trees and between rows (Fig. 1). Mulberry was selected because it grows rapidly and is notorious for the amount of damage its roots do. The zelkovas were planted as a contrasting species expected to do less damage. Containerized trees of 5-gallon size were used and were planted in square holes 30 inches on a side. For each species, eight "control" trees were planted without constraints on their roots, eight trees were planted in PVC planter boxes (hereafter called "planters"), eight were planted in wells, and eight were planted with 5-mil polyethylene around the sides of the square planting holes. Wells were 18 inches in diameter and 18 inches deep and lowered root systems 18 inches while still leaving the trunk and root crown exposed to air. Planting holes were 24 inches deep for all treatments but wells, which required planting holes 42 inches deep. 1. Presented at the annual conference of the International Society of Arboriculture in Quebec City, Canada in August 1 984. 166 Wagar: Reducing Surface Rooting In each treatment, four trees were planted adjacent to sidewalks and four away from sidewalks (Fig. 1). Among each of these sets of four trees, the planting holes or planters for two trees were filled with the native soil (a clay loam). The other two holes or planters were filled with a mixture having equal parts of sand, peat, and the native soil. For the planters, we followed the manufacturer's recommendation and use %-inch gravel to backfill between the planter and planting hole and to cover the planter. Trees were planted in July of 1980, and that fall we established a lawn of perennial ryegrass between the trees to simulate the conditions under which many city trees grow. The lawn and trees were fertilized and sprinkler irrigated each summer through 1983, providing three and a half seasons of growth. The trees reached heights of 12 to 20 feet and diameters of 2 to 5/z inches. When laying out the study area and planting the trees, we learned that parts of the field had compacted soil at various depths. These layers caused water logging in some areas, and some tree wells filled with water and remained full for days or even weeks at a time. Three of the zelkovas planted in wells died. Although the nonuniform planting site reduced the study's precision, it provided some information on the importance of soil aeration. A second difficulty, however, provided no such benefit. In rototilling to prepare the soil for seeding to grass, we managed to tear the top few inches off many of the polyethylene barriers, essentially destroying one of the treatments. Originally, we planned to spend the summer of 1984 excavating root systems. We learned, however, that high velocity dynamite can be used to excavate and study tree roots (7). Dynamite permitted excavating at the end of the 1983 growing season, providing results a year earlier without sacrificing a growing season. Because dynamiting was known to change the vertical position of roots, we excavated first with hand tools to expose and map the top 8 inches of each root system. A licensed blaster then coordinated the use of high velocity dynamite — in sticks 1 inch in diameter and 8 inches long — to loosen the remaining soil around each root system. We placed dynamite around each tree in 12 holes 30 to 36 inches deep and spaced 2 to 3 feet apart. At first we used half to two-thirds of a stick in each hole but did not get quite enough loosening. Thereafter we used 12 full sticks per tree, a little more than one stick per cubic yard of soil. All 12 sticks for a tree were detonated simultaneously. Although dynamiting worked very well except where the soil was heavily compacted, it really was not necessary. We gained almost all our information from measurements of roots exposed with hand tools in the top 8 inches of soil. We expressed the "amount" of these roots in terms of their cross-sectional area, number, and average diameter — all as measured in a doughnut-shaped zone 8 inches deep and extending from 2 to 3 feet from the center of the tree. For each tree, cross-sectional area was totalled for all roots in the doughnut-shaped zone, with each root measured at the greatest diameter attained and measured at only one point, even if branched. Only roots at least A inch in diameter were considered. Data were analyzed using the t and Bonferroni t tests with General Linear Models procedures from the Statistical Analysis System (3). The t tests establish a probability (P) that a single difference — as between means of the control and well treatments — may be due to chance. The Bonferroni t test is considerably more conservative and provides an adjusted probability (Pb) reflecting the increased likelihood, when several differences are to be tested, that one or more is due to chance (1). For cross-sectional area of roots, basal area of stems, average number of roots, and average diameter of roots, respectively, six pre-planned comparisons were made (Table 1). A t test was