We examined the impact of soil water availability and heat stress on gross primary productivity (GEP), ecosystem respiration (RE), net ecosystem productivity (NEP) and evapotranspiration in an age-sequence (71, 36, 20 and 8 years old) of planted temperate pine forests, north of Lake Erie in southern Ontario, Canada using seven years (2003-2009) of eddy covariance flux, meteorological and edaphic data. Throughout the study period, three distinct extreme weather events occurred, allowing us to study the forest response: a warm and dry spring of 2005, an extremely wet summer of 2006 and a summer drought in 2007. In 2005, NEP of the 71-year-old stand was reduced to 36 g C m -2 y-1, a 74% reduction as compared to its seven year mean value of 136 g C m -2 y-1, primarily because of a decrease in photosynthetic activity due to an early spring drought. A similar decrease in NEP was observed at the 36-year-old site, while the 8-year-old site became a large source of carbon. The 20-year-old stand, which had a high water table due to its topography, was least affected. In contrast, 2006 was the most productive year, with the highest-ever recorded GEP value of 1468 g C m -2 y-1 in the 71-year-old stand, however, hot summer temperatures also resulted in high RE (1292 g C m -2 y-1 which was ~100 g C m -2 greater as compared to the seven-year mean value), effectively making 2006 an average year in terms of annual NEP. In 2007, a summer drought caused GEP to decline but colder temperatures in late summer and autumn also reduced RE, resulting in NEP similar to its seven -year mean value. Of the seven years on record, the most productive years were 2003 and 2008, when low temperature and RE resulted in about 200 g C m -2 y-1 of NEP. An experimental study conducted in 2009 at the 71-year-old stand to evaluate the response of the forest to induced drought (270 mm, or 27% reduction in precipitation) in the early growing season showed a 48% decrease in cumulative growth rate due to a 30% decrease in normalized values of measured transpiration from the trees in a 20m x 20m drought plot compared to a reference plot. In addition, continuous soil CO2 efflux measurements using automated chambers in both the reference and drought plots showed that soil CO2 efflux from the drought plot decreased much more rapidly as compared to reference plot (i.e. 9.93 and 4.90 µmol CO2 g C m -2 s-1, respectively). Our analysis indicated a moderate and comparable suppression of both GEP and RE with increasing shallow soil water deficits below a selected threshold across all years and sites. Within this trend, however, substantial discrepancies may exist among sites and between years. Our findings highlight the importance of incorporating deep soil-water availability with increasing stand age, in addition to shallow soil moisture trends. Our study suggests that the simultaneous occurrence of early growing season drought and extreme summer heat events may play a large role in reducing the annual net carbon uptake in mature and young forests in eastern North America.