One issue with these portraits of the little girl and the previous portraits of the subject standing in the flowers with the backlight is that there is a significant blue cast - this can be addressed with a white balance correction. Although it is a little more difficult to correct an 8bpc image grabbed from the internet compared to a raw image file, I have attached an image that demonstrates the cast. In the "corrected" image, I simply made a duplicate layer in PS and invoked the Camera Raw filter and applied an auto white balance operation in it. I then changed the blend mode of this layer to Color to swap the adjusted color with the original color. I then added a curves layer to brighten the overall tone in the image.
The white dot depicted in the center of the subject's forehead is a reference for the skin tone - it is an area that usually does not get flush or rosy with cooler temperature and, in adults, is not a heavily made up area with potential artificial colors applied. As can be seen, the a and b values for the area in the before and after images demonstrate a significant blue cast (the b value is negative, denoted by parentheses, in the original and positive in the after image). Skin tone, regardless of gender and racial differences, has a pretty consistent composition of positive a and positive b, with b typically slightly greater than a. This is a good rule of thumb for checking your skin tones "by the numbers" so that an inaccurate display profile or tired, adapted eyes do not interfere with color editing.
If you are shooting in open shade, then it is not surprising that there is a blue cast - you are getting primarily skylight illuminating your subject, as opposed to direct sunlight. Add to that shooting in dappled shade under a green leafy canopy, which can introduce a green cast into the scene, and you get blue-green illumination - one might counter this by adding the opposite, which is magenta, making things that are already way too blue turn purple. The original area on the subject's forehead, reading a16 and b(7) is purplish, which is not a typical skin tone and is visually disconcerting. Notice that the after image retains the a16, but adjusts the b from (7) to 20, where it should be and the entire scene falls into place.
I hope this is helpful. As an aside, along the lines of Kim's post, I have deleted the file from my computer that I used to generate the comparison image. I exercised my judgement to use the image for explanatory purposes, but I also want to respect your personal work and the privacy of the subject in the image. Please post here if you would prefer that I remove the attached image from this post as well.
Kirk
EDIT - If you are unfamiliar with Lab color jargon: "a" and "b" are color axes in the Lab color model. a represents the green-magenta axis and b represents the blue-yellow axis. Negative a values tend toward green, positive a values tend toward magenta (red). Negative b values tend toward blue, positive b values tend toward yellow. You can think of the a axis as the "tint" slider and the b axis as the color temperature slider in a typical CCT white balance adjustment model like that in ACR/LR.
a16 b(7) = +a and -b = bluish red = purplish
a16 b20 = +a and +b = reddish yellow = orangish, or skin tone.
To address a white balance issue you can shoot in AWB, set a specific, preset WB in camera, or shoot a WB reference card or target in the scene lighting and set a custom white balance in camera. Of course with a raw file you can always adjust WB during conversion, but you still need a neutral reference if the lighting is not straightforward.
If, as you mentioned, you printed this image and it looks good, then maybe this color palette is simply what you prefer and white balance is simply a preference and not a literal rule for the images you like to make.
Image hosted by forum (
1062629)
© kirkt [SHARE LINK] THIS IS A LOW QUALITY PREVIEW. Please log in to see the good quality stuff. 
