Terraces at the author’s Khadighar Farm hold nutrients and water while supporting crop growth on a slope. |
By Will Bonsall
Photos and illustration by Will Bonsall
Around the world people have used terraces since ancient times to grow crops on steep hillsides. The benefit of preventing soil erosion is obvious, but an equal value is reduced runoff of soluble minerals with rain.
Flat land has the opposite challenge: Excess water tends to sit on overly flat land too long; having no place to go, it can plug soil pores and exclude air. A form of terracing can remedy that as well.
Terracing benefits both flat and steep land by concentrating all the slope into extra-steep banks, or even vertical walls, and creating level, well drained ground between. That can make the steep divider-strips much more prone to erosion, so they’re usually allowed to sod over in order to shed water harmlessly. These sodded banks, called “lynchets,” are lost to food production, but the increase in crop yield from the rest of the area, largely due to better water and nutrient relations, more than compensates. In fact, terracing is the only way really steep slopes can be farmed sustainably.
Those lynchets, while not tillable, are not unproductive. I mow mine at least twice a season and use the grass for mulch and compost. The grass yield of those strips is proportionately greater than that of my open hayfield, since they are adjacent to the well-fertilized terraces. Moreover the sod traps and sequesters fugitive nutrients and returns them to the cropland as mulch or compost.
Terracing steep marginal lands requires lots of labor to shape them, so people have resorted to such intensive strategies usually in areas with dense populations and a shortage of more convenient land. But what about land that is not so steep, where erosion is not so glaring and agriculture appears to be more sustainable as is? Is there no benefit to terracing those areas? Gently sloping land may be less prone to visible erosion, but soluble minerals are still carried away in runoff. The remedy for that is to minimize runoff at the source. The more drops of rainfall or snowmelt that leave the plot by percolating downward (as opposed to running off), the more soluble and suspended minerals are trapped among the soil particles, each of which has a micro-electric charge, a “magnet” for free cations (positively charged ions). Better yet, some of those dissolved minerals become entrapped in humus, which holds them until subsequent crop plants appropriated them. The water itself spends more time in the subsoil, within reach of plant roots. Therefore even a gentle slope can be greatly improved by terracing, and with less effort than steeper slopes.
Is it worth it? I’ve found the benefits are profound and long-lasting – a definition of sustainability. But how to go about it – and how do you terrace land that is level to begin with? Let’s look at that question first.
Sculpting the earth into terraces involves moving soil from one part (uphill) to another (downhill). If the land is already too flat to have any uphill or downhill, we can create some by digging deep ditches and spreading the removed soil evenly on the rest of the plot, especially on any low spots. The combination of lowering the low point – the ditch – by, say, a foot or two, while raising the plot surface an inch or two effectively creates a terrace, a flat area where water sinks right in yet is eventually drained away slowly by the ditches.
A flat terrace surrounded by more flat land almost defies our definition of a terrace; maybe it’s better to think of it as a giant raised bed; the effect is largely the same. When the early Aztecs were forced by their enemies into the swampy marshes of Lake Texcoco, they created usable cropland by digging canals through the wetlands and piling the rich lakebed mud onto the remaining shallows, forming “chinampas,” or what Europeans called “floating gardens.”
When laying out terraces on hillsides, we want them to serve forever, so they deserve a bit of thought. Ideally terraces should follow the contour of the slope; if they don’t, then a lot of energy must be expended to move dirt from one end to the other. The Oriental rice terraces we see in National Geographic pictures typically curve gracefully around the hillside and are often wider at some points than at others due to variations in grade. (Steeper grades require narrower terraces in order to maintain the same height.) In different parts of my land I’ve used different strategies.
Two of the five terraces in the Lower Garden. Arrows indicate the direction of flow of water. Each terrace is 15 feet wide to accommodate two intensive, 40-foot-long beds, two beds wide, and walking paths. A cross path separates the two 80-foot beds at 40 feet. |
As the five terraces marched down the hill, the variance from natural contour became so extreme that the last terrace is split in two crosswise, with a cross-lynchet to separate the levels. |
My first project (Lower Garden, 1970s) was on a curved slope, so I curved the terraces (see illustration). This land varied a lot in steepness – and thus width – from one end to the other, but I was determined to make the five terraces parallel – each 15 feet wide to accommodate my intensive beds, and with two 80-foot-long beds. On some terraces I had to move lots of earth from one end to the other. As the five terraces marched down the hill, the variance from natural contour became so extreme that I split the last terrace in two crosswise, with a cross-lynchet to separate the levels.
In my so-called Back Garden, I was determined to make the terraces parallel and straight, for ease in laying out rows with a stretched string. Most of that 160-foot slope was barely curved anyway, so it required minimal adjustment. At a fairly sharp curve at one point, I compromised and made an abrupt bend, continuing on straight for another two 80-foot bed lengths. This avoided a problem I had in the Lower Garden: Since I cannot stretch a string from end to end to mark rows, I must put in a number of marker stakes, say every 8 or 10 feet, and warp the string around them to follow the terrace curve. This looks more “charming” than the straight terraces but is more of a nuisance to plant, especially to grains.
Not being a surveyor with proper instruments, I improvised my equipment. At first I used a 75-foot garden hose with a clear glass tube inserted in each end (clear plastic would be better, but glass was what I had). I filled the hose with beet-juice-tinted water and corked both ends temporarily, to form a crude spirit level. A helper 75 feet away would move up and down the slope until the liquid was about to spill out (cork removed), indicating that he was at the same level as me. He drove in a 4-foot marker stake there. A number of those stakes gave me a contour base – or reference-line – from which I could move up or downhill. That worked okay, but for later projects I used a builder’s pop-level to establish contour points. For final shaping and smoothing of the finished terraces, I used a 16-foot 2×6 on edge, with a 4-foot spirit level; I frequently checked it in different directions to minimize accumulated error.
To actually form the initial terraces, I used a tractor and moldboard plough. My neighbor, whose ancestors had farmed that land since 1803, for better or worse, pointed out that when you plow “lands” (or sections of a large piece), the first pass lifts out a sod and lays it over onto the unbroken sod, leaving a furrow; henceforth each sod is turned into the previous furrow. When the piece is all plowed, you still have the slight hump (called a “live furrow”) on the beginning side and a final depression (a “dead furrow”) on the finishing side. Subsequent harrowing pretty much smooths the piece. Moreover the plowman usually starts and stops in a slightly different place every year to avoid building up a ridge and ditch. But, suggested my neighbor, what if you deliberately started and stopped at the same point each time, placing the downhill tire on top of the live furrow? If you did it several times and on narrower strips, you would be constantly building up one side while scooping out the other side, and if your plowing followed the contour, you would begin forming a terrace.
I used this method to save countless hours of hand labor, in effect using the tractor and plow like a bulldozer, which I do not own. One could do the whole project this way, especially if the terraces aren’t very high, but at some point it becomes impossible: Each time the downhill side (live furrow) gets higher and the uphill side (the dead furrow) gets deeper, so setting the uphill wheel on the next terrace edge presents a great danger of tipping the tractor (see drawing). Even if this weren’t a problem, at some point the plowshare ceases to reach the ground and you’re plowing sky. At that point you must park the tractor and shift over to spade and wheelbarrow – and pry bar, if your land is anything like mine. The outer half of each terrace is now double-deep, beautiful loose topsoil; the inner half is scalped primordial glacial drift, rich in minerals and devoid of organic matter, very hard and very stony. I typically spend three or four years of green manuring, while adding mega-doses of forest residues (which is how Nature creates most new soils). Eventually my entire terrace becomes equally fertile, as my crash restoration program mimics several centuries of natural soil building, and I can grow any crop on any part of it without fear of erosion. Sustainability is for the long term, and that’s what these terraces are about.
What determines their size? That depends mainly on the natural grade and how much earth you’re willing to move around. A larger number of narrower terraces will take less labor than fewer but wider ones. But there are problems with that: Since each terrace must have an access path on the rim as well as on the inner side (on flat land you would only need one of them), more space is wasted. For example, all my cropland is laid out in 6-foot-wide beds (including the path), so a two-bed-wide terrace must be 12 feet wide plus about 2 feet for the outer path. Therefore the more beds I have sharing that extra path, the less hard-gained land I devote to footprints.
Another problem with narrower terraces is that each of those sod-covered lynchets is an “edge-zone,” a frontier from which grasses and weeds try to encroach upon the cultivated land. This is an issue with any garden not surrounded by pavement, but why increase the work by creating an unnecessary amount of perimeter?
Many “ornamental” terraces are retained by landscape timbers (pressure-treated with toxic stuff, in case you care) or stone walls. These don’t require mowing, but they also don’t produce anything useful. They might seem to save space, since the retaining surface is vertical as opposed to the 45-degree slopes of my sod lynchets, but that advantage is misleading, since any retaining wall not backfilled with a foot or more of rubble or stones will soon succumb to the inexorable power of frost. That rubble not only produces no crops but is constantly being churned up when you till near it – there’s no precise boundary. We have one such terrace – the perennial flower garden – and the reason that works is that the crest of the stone wall is overrun by creeping plants, which need little hand-weeding and no deep digging.
Whether terracing makes sense depends on your timeframe. If you plan to use this garden space for only a few years, it probably isn’t worth it. If you intend to eat from this land, and enable others to do so forever, then terracing makes all the sense in the world. Your considerable initial investment in labor will pay you and others rich dividends for many decades and, perhaps, centuries.
About the author: Will Bonsall lives in Industry, Maine, where he directs Scatterseed Project, a seed-saving enterprise. He is the author of “Will Bonsall’s Essential Guide to Radical Self-Reliant Gardening” (Chelsea Green, 2015). You can contact Will at [email protected].