One lesson we learned at this year's annual Annan Salmon School happened whilst we were all asleep. We had gathered on the banks of this classy river on the Sunday. The river was high after a bout of rain, but clearing and falling nicely and, as casting instructor Glyn Freeman put the tyros through their paces – first with a single-hander, and then with the double-hander – so I watched fish splash all through the pool, with a high degree of anticipation. Being Sunday, and being Scotland, meant that we couldn't fish for them, but never mind ... there was always tomorrow.

Unfortunately, as sure as Monday occurs after Sunday, it poured down. All night. So, the next morning, with the weather now fine once again, as Glyn and I polished up half of the group's Spey-casting at Kirkwood, we watched the river rise, and rise. Not good for salmon fishing at all; we want a falling river. But how long would it rise for? We were kept informed by text messages from Anthony Steel, who was guiding the other half of the group further upriver at Jardine Hall. Using sticks planted at the river's edge, he could plot the river's rise. "Midday, still rising"; "Still rising at 1pm, but more slowly". Finally, he texted that the flood there had peaked at 2pm. Now his group would start fishing in earnest, making the most of a falling river, but how long would it be before it peaked at Kirkwood, seven miles further downstream? The answer was 4.30 pm. Just about when it was too dark to fish. Still, it was a good opportunity to get a day's extra casting practice.

We're always learning on the salmon school, and this Sunday night downpour was revealing. It enabled us – through our own, simultaneous observations on two beats, plus the facility to access the river-level gauges via the internet – to determine just how long after rain the pulse of flow will peak at the gauging stations, and also on the respective beats we fish. There are two gauge stations on the Annan. The key one for us was the gauge at Three Waters Meet – high upstream, near Moffat. This, we later learned from the internet, peaked at midday, nearly two hours before the river peaked at Jardine Hall, 12 miles further downstream. Yet it took further two-and-a-half hours for the river to peak at Kirkwood, another six miles downstream. This is due to basic hydrology, because the river is bigger here, and has tributaries of both Dryfe Water and Kinnel Water add to its volume. The level does not peak all over the river at the same time, and the peak itself can take an inordinately long time to migrate downstream.
The lesson here for anglers is that after a bout of rain, there is probably fishable, falling water at the very top of the system after just a few hours, whereas far downstream the river will be continuing to rise and could be unfishable for most of the day. Thus, the flexible, mobile angler can make the most of conditions by migrating upstream, whereas the angler restricted to fishing one lower beat is forced to wait. In this instance, in the case of Kirkwood's anglers, it took all day. Never mind, they were all really good at Spey-casting by the end of it!