diff --git a/instruments/readers/nasa_ames_reader.py b/instruments/readers/nasa_ames_reader.py
index 1601542..5c71f68 100644
--- a/instruments/readers/nasa_ames_reader.py
+++ b/instruments/readers/nasa_ames_reader.py
@@ -75,6 +75,22 @@ def extract_var_descriptions(part2):
 
     return descriptions
 
+def generate_missing_value_code(max_val, num_decimals):
+    """
+    Generates a missing value code consisting of all 9s.
+    - `max_val`: Largest expected valid value in the column.
+    - `num_decimals`: Number of decimal places to preserve.
+    """
+    # Determine order of magnitude (1-2 orders larger than max value)
+    order = int(np.floor(np.log10(max_val))) + 2 if max_val > 0 else 2
+
+    # Construct the missing value code as all 9s
+    if num_decimals > 0:
+        missing_code = float(f"{'9' * (order + num_decimals)}.{ '9' * num_decimals }")
+    else:
+        missing_code = int('9' * order)
+    
+    return missing_code
 
 
 
@@ -157,6 +173,11 @@ def read_nasa_ames_as_dict(filename, instruments_dir: str = None, work_with_copy
         if 'end_time' in df.columns:
             df['end_time'] = df['end_time'].apply(lambda x: start_date + timedelta(days=x))
 
+        variable_missing_values = []
+        for col in df.columns:
+            missing_value = generate_missing_value_code(df[col].max(skipna=True), 4)
+            variable_missing_values.append(missing_value)            
+
         # Create header metadata dictionary
         header_metadata_dict = {
             'header_length': header_length,
@@ -164,6 +185,7 @@ def read_nasa_ames_as_dict(filename, instruments_dir: str = None, work_with_copy
             #'num_dep_vars': num_dep_vars,
             'variable_names': vars_list,
             'variable_descriptions' : var_descriptions,
+            'variable_missing_values': variable_missing_values,
             'raw_header_part1': part1,
             'raw_header_part2': part2,
             'raw_header_part3': part3