While I wait for Mother Nature to inflict her next cruel deed on the garden, I’ve been busy planning new plantings. It depends on the garden category: new gardens, small gardens, large gardens in need of regeneration, large gardens that are spectacularly empty and need to be replanted throughout. Planting plans can be micro or macro. You can be obsessive, down to the last mini-primrose, or you can be relaxed about which plants in the big bunch will end up where. It is interesting to see what type of planner a new garden owner is at first and what kind of planner he becomes as he gains experience.
When the planting plan actually has to be planted, I leave myself a degree of freedom, but it’s a degree of freedom framed by a broader plan of what needs to go where. Can planting plans just be done digitally, especially with AI? For me, no, I’m not digitally savvy and this time of year I get screen shock after dealing with online tax returns. Masu. Add-ins compiled with the VAT system did not work, and the Tax Office Compatibility Hub did not work. Failed on New Year’s Eve.
In your case, the answer is probably yes. I am discussing a planting plan, not a garden plan without plants. Digital software and systems are as useful as spades, but their existence has not eliminated the need for human skills in planning planting.
Vegetable gardens are simpler and have limited acreage, making them better suited to technology. The content of flowerbeds is more complex and often less formal. Search engines and apps like iscapeit.com will give you a list of flowering plants of various heights, colors, and seasons, but those lists anyway exist in catalogs and books. Asking targeted questions of AI resources (aigardenplanner.com is one of them) can also help planners find plants with graduated colors and heights, but they can also help you come up with great personal choices. There is no such thing. For example, a tall leaf blade in the middle of the length of the border or a fallen specimen. Aster frikartii Mönch spaced along the frontier of the border. Its unusual height is difficult for technical sources to weave together.
The digital program can create a scaling plan that can be printed on graph paper, spacing each plant as required. Digitally savvy beginners may initially find this plan reassuring. Admittedly, this plan is better. Is it a break from the “bad” old days and bypassing them with a mass of sketches and ink? So far, I disagree.
Some content could not be loaded. Please check your internet connection or browser settings.
This printed version is essentially the same as the handwritten planting plan devised by garden guru Miss Jekyll some 125 years ago. Rather than circular clumps, she painted plants intertwined in a stream, sometimes micromanaged. She also specified the exact number of plants of each variety. Her plans were ignored after her death, but not because technology made them unnecessary. Many of them were left to rot in a shed in Somerset. They were purchased from indifferent Britain and taken to the United States.
Technology tools and digital resources depersonalize processes. Would using them take away the serendipity, the happy event of coming up with a random choice and having it work out really well?I think that took it a step back. Applying new searches to AI could make serendipity appear sooner. Certainly the programming will improve, but we wait to see if future Hortobots can plan in the style of, say, Sissinghurst Gardens and its founder, Vita Sackville-West. Some of Sissinghurst’s most famous plantings are the work of her successors at the National Trust, particularly Graham Thomas, and the genius duo she left in charge, Pamela Schwerdt and Sybil Kreutzberger. This plan will fail because of this. Hortobot would have to comb through some 12 years’ worth of online records of Sackville West’s gardening columns to look for genuine evidence of her style.
A frightening prospect looms. Could Hortobot, or even ChatGPT, come up with a Robin Lane Fox-like planting plan based on an even larger database, my weekly FT column, which turns 55 this month? But writings, even those of great gardeners like Sackville West, change over time. Will sifting through them, ignoring dates, produce a genuine plan like the one the author endorses?
Digital garden plan created using Planner 5D
Between planning and actual results, real-world, digitally derived planting plans are at risk. Worship can crumble if the bit fails in reality or becomes cloudy, sluggish, or dry after hitting the ground.
The problem of space is compounded by the problem of time. What types of gardens does the planting plan apply to? Trying to enhance an existing plan by digital sources is difficult. Does over-specifying the shape and assets of existing plants limit the range of what appears to be good new companions?
In large empty gardens, I often place one or two plants in different locations in planters on the fly. Over-prescribing in advance destroys spontaneity. New technology waves may try to incorporate that, but I think that’s the opposite of how technological systems work.
We’re waiting to see if the Hortobots of the future can plan something like, say, the gardens at Sissinghurst.
So here are some hard-won pre-digital principles for each type of garden you plant. Before selecting and placing shrubs and trees, pay attention to their height and width. We all plant too closely together, a mistake that is compounded by our growing impatience. While you may be tempted to move before this mistake becomes too serious, it’s still a bad mistake.
Small enclosed gardens can include height and extent to a degree that is difficult to account for with digital sources. Pruning beyond book and list recommendations makes it possible to include some giant plants. In an enclosed space, one or two extra tall features work well and can add a jungle-like look. To do this, choose upright plants that don’t have too many leaves.
In large gardens, this pruning and increasing extra height looks wrong. In it, space the shrubs to their mature dimensions and fill the spaces between them with temporary fast-growing covers. Low-growing buddleja or ceanothus varieties are good, as are pink-white mallow or labella shrubs that can be passed between them.
Recommended
Note that in small gardens, a mature wall shrub will dry out the soil beneath it and around its spreading roots. Most members of the Clematis family are an ideal replacement, as they prefer to emerge from other plants, as long as their roots are regularly watered and nourished. If you want to line up trained trees to block out neighboring trees, keep in mind that as the tree’s roots get older, they will also block the plantings underneath.
One of my inspirations for gardens is Helen Dillon, famous for the sublime gardens she devised in Dublin during her working days. I first heard her story in 1993 when she was interviewed on Irish prime time television. She was asked what she thought about the garden plans. “That’s a hell of a plan,” she retorted. It’s not because her garden doesn’t have a clear plan, but because its planting evolves as a process, with one thing suggesting another to her ever-vigilant eye. AI and digital resources could not have achieved that. It will probably never be achieved.
Check out the latest stories — follow @ft_houseandhome on Instagram
