A change is gonna come: Anna Pavord's plans for the year ahead in her Diary of a Dorset Garden

A garden is a process, not a product. It never stays still. Nor should we ever expect it to. Part of the process has to do with the way our own lives change. At different times, we need different things from the space outside. For what seemed like endless years, our garden was a loosely connected series of spaces round a swing, sandpit, gymnastic display area, bicycle heap, guinea pig run, rabbit hutch, sheep tether, paddling pool and swingball. Imperceptibly, though, the landscape changes. The sandpit becomes a pond, the swing is replaced by an apple tree, the livestock dies and is buried with extravagant pomp. The possibility of growing some decent grass at last arises.

Inability to see a garden as anything but a product was the central problem with the terrible makeover programmes that for a time dominated our screens: gardens were delivered fully formed to bemused owners who had neither been involved in their making nor briefed about their maintenance. Up to a point, the approach works with houses, because a room, once "done" stays in place until we get tired of the scheme we've chosen. At that point, no amount of new cushions will satisfy the itch to do something different.

Even in the relatively short time we've been in the garden I still think of as "new", there have been changes, though mostly they've been thrust upon us rather than instigated. Things die. A weird disease has been nibbling away at the willows that do a good job stabilising the very steep bank that ends at our bottom boundary. When we first came, the trees were too tall to take the wind that funnels up the valley. Willow wood is brittle and branches kept snapping. We pollarded them and they resprouted enthusiastically. I like pollards and thought this was, in every way, a change for the better. Pollards are easy to manage. Every winter you cut off shoots breaking from the trunk below the pollarded head and take out one or two of the biggest branches from the head itself. Sustainability incarnate.

Then branches began to die back. At first, this didn't matter, as we were cutting out some each year anyway. But the losses accelerated and the bark seemed almost stripped from the wood. At first we thought it might be squirrels. Now I think it's more likely to be one of the fungal diseases that attack willow: anthracnose (Marssonina salicicola) or scab (Fusicladium saliciperdum) perhaps. Either can cause lesions which may girdle a branch and so cause it to die. Fortunately the willows weren't special ones, and so when several died completely, I wasn't too sad.

But just in that one area of the garden, the two changes have had a big effect. First, the pollarding, which meant that for the first time we had a clear view over the tops of the pollards to the string of alders beyond. Then the death, which immediately in any gardener's mind brings up the question of replacement. Or not. After the trauma of the Great Storm in 1987, quite a few gardeners eventually came to the conclusion that the loss of a tree was not necessarily a bad thing.

Here, though, we needed tree roots to help hold the steep ground in place. And the sudden gap in the screen that fills the whole of the rest of the bank looked weird. Because the land just the other side of our boundary belongs to the Dorset Wildlife Trust, we needed trees that would shake hands, as it were, with the natural landscape beyond. Most of our boundary bank is filled with holly trees and stooled hazel, both of which I love. At the far end is a single rowan (Sorbus aucuparia), stripped now of all its berries, but a handsome, useful tree, fairly upright in habit and even when mature, scarcely taller than seven metres. It would never need pruning or lopping, which on a bank this steep, is a consideration. So gradually I played around with the idea of planting a couple more rowans to replace the dead willows.

After testing them against various other possibilities – spindles? geans? – the rowans seemed the best option. And since the place was fit for them (they like damp ground), they are now planted. Next year, there will be both blossom and berries in a place where neither has been before. Rowan, or mountain ash (depending on where you were brought up) brings a lot of baggage with it. Of all trees, it is the most powerful at protecting against witches.

So even though this was a change that we didn't initiate, in the end I think we'll benefit from it. But of course things grow, as well as die and the change that this brings about in a garden is more imperceptible. Shade is cast where there wasn't shade before. You find yourself bobbing and weaving on paths to avoid branches that are now overhanging or thrusting out much faster than you anticipated. I can scarcely bring myself to snip bits off our Magnolia delavayi but it has to be done. It moved here with us only a few years ago and at that stage was scarcely 45cm tall. It took a long time to decide whether it was going to live or die. Having decided, it's now growing at a phenomenal rate, more vigorous though on the side where it adjoins the path than on the other, where it could stretch as far as it wanted.

So I snip, saying "Sorry, sorry, SORRY" as I do it, because after all it was me that put it there in the first place. But at least you can make the whittling as seamless as possible by cutting a branch only where it meets the next, bigger one. The worst thing to do is to barber a stem halfway down its length. It may do the job, but it shows all too plainly that the job has been done. Anyway, my New Year's resolution is Think Ahead.