The last 10 years are far too short a time period to establish anything due to the variability of the data. If you include more data you get a more accurate res

The last 10 years are far too short a time period to establish anything due to the variability of the data. If you include more data you get a more accurate result, so do that. 30 years is the usual minimum period considered for climate, so what's your purpose in choosing something considerably shorter? Edit: That plotting page is quite nice and illustrates my point well. Plot the years 1987 - 1996. Notice that the trend for that decade is actually down. And yet, add another 10 years to take it to 1987-2006 and it gets pretty close to the 30 year trend.