The question, asked properly
"Does region change taste" is nearly always a comparison in disguise: Darjeeling tastes unlike Assam, Uji unlike Kagoshima, Wuyi unlike Anxi. All of that is true and none of it establishes what people take from it. The reason is that when you move between two tea regions, at least four things change at once — the PLANT, the CLIMATE, the PROCESSING and the PICKING STANDARD — and only one of them is what terroir is supposed to mean. Assam and Darjeeling differ in variety (assamica against largely Chinese-variety and clonal), in altitude, in manufacture, and in what a picker takes. Attributing the whole difference to the hillside is an attribution with three unexamined confounders in it.
What place genuinely contributes
TeaHQ's own reading of well-attested general knowledge, with no single source behind it.
Three effects are well supported and worth naming precisely. GROWTH RATE. Cooler conditions — from altitude, latitude or season — slow shoot growth. Slower growth concentrates what accumulates in the leaf and changes the ratio of the compounds that read as sweetness and body against those that read as briskness. This is a real mechanism and it is the honest core of most altitude claims. WATER AND LIGHT. Rainfall distribution, cloud cover and mist affect the plant's growing conditions continuously. The famous mist of several high districts is doing something close to what a shade cloth does deliberately in Japan, and shading has measurable effects on leaf composition. SOIL, WHICH IS THE WEAKEST OF THE THREE. Tea wants acidic, well-drained soil and gets it nearly everywhere it is grown commercially. Specific mineral transfer from a particular rock into a particular flavour is asserted constantly in tea marketing — the yancha "rock rhyme" is the best-known case — and the evidence for a direct causal path is far thinner than the confidence of the claim. Something real is being described; the mechanism usually offered for it is not established.
The test that separates place from everything else
Hold the cultivar and the process constant and move the location. This happens more often than you would think and the results are informative. The same Japanese cultivar grown in Uji, Shizuoka and Kagoshima, steamed and rolled the same way, gives noticeably different tea — Kagoshima's warmer, earlier season against Uji's cooler one. That is a place effect with the confounders removed, and it is real but smaller than the difference between two cultivars in the same field. Darjeeling material planted in Nepal, on continuous terrain, produces tea that experienced buyers frequently cannot separate blind from Darjeeling. That is close to a controlled experiment and its result is inconvenient for the strongest version of the terroir claim. Taiwanese oolong methods taken to Vietnam and Thailand by Taiwanese producers give a recognisably Taiwanese style of tea. Process travels; the mountain does not.
So what is a region name actually telling you?
Usually a package of four things sold under one word: a set of cultivars that district plants, a processing tradition that district practises, a climate and elevation band, and a picking standard. That package is genuinely predictive — a reader buying "Anxi Tie Guan Yin" can expect a lot about the cup — and it is predictive because of the whole package rather than because of the ground. That is a better claim than the mystical one, not a weaker one. It explains why regional character is real, why it can be partially reproduced elsewhere, why it changes when a district switches cultivars or roasting fashion, and why a region name plus a year plus a producer tells you far more than a region name alone.
Where the claim is used to sell something
Two shapes are worth recognising. The first is a region name applied to material from outside it, which is common enough that several districts have geographical-indication enforcement and litigation behind them. The second is a mechanism asserted with confidence and no support — soil minerals passing into flavour, a specific rock imparting a specific note. Neither means regional character is invented. It means the label is doing two jobs at once: describing a real package of plant, process and climate, and carrying a premium that the package alone might not command.