The STIS NUV MAMA detector displayed a stable and low dark rate before August 2004, at a level of ~1.1E-3 count/sec/pixel. During this period, after a shutdown the dark rate would spike to a slightly higher level and decline within weeks. When STIS was restarted in August 2009, it was found that the spike in dark rate was more than tenfold higher than anticipated. This rise is believed to be the result of a prolonged period at cold temperature, that led to a buildup of a meta-stable population of impurities in the detector. In December 2009, a pair of new dark and TDC (temperature-dependent dark correction table) files were delivered, to provide a base of dark subtraction for post-SM4 data. However, in 2010 the STIS NUV dark rate continued to decrease, leading to uncertainties in the dark subtraction of ~30% for new data taken between January and August 2010. To make the situation more complicated, STIS was placed in a safe mode between August and November of 2010, resulting in another surge and decline in its NUV dark rate. An updated TDC file (v2a1955fo_tdc.fits), which provides a set of flux normalization scales at representative dates was delivered to CDBS on February 10,2011. A verification process with the dark images themselves demonstrates that CALSTIS can now reduce the dark term to a level below 5%, except for 2-3 data points. Since December 2010, the NUV dark rate has been declining very slowly. We anticipate that this latest TDC file will provide a reasonable dark subtraction to a 5% accuracy for data taken in 2010 and 10-20% accuracy throughout 2011.