Subaru observations of A1689 ( z = 0.183 ) are used to derive an accurate , model-independent mass profile for the entire cluster , r \lower 2.15 pt \hbox { $ \buildrel < \over { \sim } $ } 2 { Mpc } / h , by combining magnification bias and distortion measurements .
The projected mass profile steepens quickly with increasing radius , falling away to zero at r \sim 1.0 { Mpc } / h , well short of the anticipated virial radius .
Our profile accurately matches onto the inner profile , r \lower 2.15 pt \hbox { $ \buildrel < \over { \sim } $ } 200 { kpc } / h , derived from deep HST/ACS images .
The combined ACS and Subaru information is well fitted by an NFW profile with virial mass , ( 1.93 \pm 0.20 ) \times 10 ^ { 15 } M _ { \odot } , and surprisingly high concentration , c _ { vir } = 13.7 ^ { +1.4 } _ { -1.1 } , significantly larger than theoretically expected ( c _ { vir } \simeq 4 ) , corresponding to a relatively steep overall profile .
A slightly better fit is achieved with a steep power-law model , d \log \Sigma ( \theta ) / d \log \theta \simeq - 3 , with a core \theta _ { c } \simeq 1. ^ { \prime } 7 ( r _ { c } \simeq 210 { kpc } / h ) , whereas an isothermal profile is strongly rejected .
These results are based on a reliable sample of background galaxies selected to be redder than the cluster E/S0 sequence .
By including the faint blue galaxy population a much smaller distortion signal is found , demonstrating that blue cluster members significantly dilute the true signal for r \lower 2.15 pt \hbox { $ \buildrel < \over { \sim } $ } 400 { kpc } / h .
This contamination is likely to affect most weak lensing results to date .