[R-sig-ME] random effect NaN corr interpretation?

[Mollie Brooks]

Hello,
I fit an LMM and need help understanding the results. They may indicate that I'm overfitting, but I have many observations (37553). I think that different Tags (random) may respond differently to the fixed effect I(SHTS*HT) so I fit a random intercept and slope for each Tag. In the model with only a random intercept, the Variance is estimated to be 0 by lmer. In this more complex model with both a random intercept and slope, the variance of the intercept is still 0 and the variance of the slope is 5.7392e-03. The part I'm not sure about is that their correlation is NaN. How do I interpret this?
thanks,
Mollie
> sm1.3=lmer(grow1~I(SHTS*HT)+SIZE+(I(SHTS*HT)|Tag_PLOT)+(1|year)+(1|PLOT), data=sHlong)
> summary(sm1.3)
Linear mixed model fit by REML 
Formula: grow1 ~ I(SHTS * HT) + SIZE + (I(SHTS * HT) | Tag_PLOT) + (1 |      year) + (1 | PLOT) 
   Data: sHlong 
    AIC    BIC  logLik deviance REMLdev
 448411 448488 -224197   448394  448393
Random effects:
 Groups   Name         Variance   Std.Dev.  Corr  
 Tag_PLOT (Intercept)  0.0000e+00  0.000000       
          I(SHTS * HT) 5.7392e-03  0.075758   NaN 
 year     (Intercept)  2.6760e+02 16.358366       
 PLOT     (Intercept)  3.1679e+01  5.628452       
 Residual              8.7827e+03 93.716081       
Number of obs: 37553, groups: Tag_PLOT, 5672; year, 11; PLOT, 10

Fixed effects:
              Estimate Std. Error t value
(Intercept)  28.823909   5.517252    5.22
I(SHTS * HT) -0.211216   0.004152  -50.88
SIZE1-ha     -1.832286   3.866764   -0.47

Correlation of Fixed Effects:
           (Intr) I(SH*H
I(SHTS*HT) -0.075       
SIZE1-ha   -0.279  0.012
> attr(VarCorr(sm1.3)$Tag_PLOT,"correlation")
             (Intercept) I(SHTS * HT)
(Intercept)            1          NaN
I(SHTS * HT)         NaN            1