ORDER AND MEMORANDUM OF DECISION
Subsequent to the filing of the complaint in this case (18 May 1981) Dr. Arnold Banner, a Ph.D. in marine biology employed by the Fish and Wildlife Service, conducted an extensive investigation into the boundaries of suspected wetlands on the defendants’ property, the nature of the wetlands and the loss to the surrounding ecology caused by filling operations on the property by the defendant William Lambert. Dr. Banner studied the aerial photographs in evidence as Government’s exhibits 4A, B, C and D. He studied topographical surveys of the *369property, compared measurements of water levels at the culvert south of the defendants’ southwest corner with water levels in the Banana River at the Port Canaveral locks. He examined the vegetation on the defendants’ property. For that purpose he made several site visits before the litigation commenced and a total of nineteen between 1980 and 1984.
Government exhibit 6 shows a transect line along which Dr. Banner caused elevations to be measured. Overlays to Government exhibit 6 contain photographs taken by Dr. Banner during his site inspections from 1980 to 1982. Many of these photographs show various species of plants relied upon by Dr. Banner as indicators of wetlands. Such plants are listed by the Army Corps manual in evidence as Government exhibit 110 as indicators of wetlands because of their tolerance for life in saturated soil. The plants found by Dr. Banner through his photographic and on site investigations are listed on Government exhibit 12. These plants were, according to Dr. Banner, prevalent in the areas which he ultimately concluded to be wetlands. Based on Government exhibit 5, 8A and 8B, he concluded that the elevation in the suspected wetlands before the filling operation began ranged from .75 feet to 1.5 feet above mean sea level or an average of .9 feet above mean sea level.*
Dr. Banner testified to a salinity gradient as follows: sea water — 35 parts per 1000; Banana River water — 30 parts per 1000; tidal creek south of Lambert property — 22 parts per 1000; station 10 in western wetlands (Government exhibit 5) 14 parts per 1000. He stated this gradient is typical of an estuary. Dr. Banner personally observed on numerous occasions that when water was measured at 1.5 feet NGVD at the culvert southwest of the property, the water would flow north through the culvert and along the ditch across the south side of the defendants’ property all the way to the southeastern end of the property.
The average water level in the Banana River according to the evidence in this case is .57 feet above NGVD. Based on a study of annual changes in water levels of the Banana River, Dr. Banner concluded the river should cause innundation on the property in its unfilled state at least one hundred days a year. This was grounded in part on water level readings gathered at the Port Canaveral locks during a ten year period. The record of water levels during the study period indicated that water levels at the Port Canaveral locks which are less than two miles from the defendants’ property equalled or exceeded .9 feet NGVD ninety-six days a year and equalled 1.1 feet NGVD fifty-seven days a year (see Government exhibit 13). Based on visual observations of the soil on the defendants’ wetlands, Dr. Banner testified that it was saturated throughout the year — as opposed to innundated.
The Government’s hydrologist, Michael Lofton, corroborated Dr. Banner’s findings. He testified that in a typical year water levels at the Port Canaveral locks would equal or exceed .9 feet NGVD one hundred twenty-five days per year. He further testified that when the water level in the Banana River equalled or exceeded 1.0 foot NGVD, the culvert southwest of the property provided a “good hydraulic connection” to the Banana Rivér. He also testified, however, that the culvert would be an obstruction to the northward flow of the water from the river at river stages less than 1.0 foot NGVD. His study of the water data taken from the Port Canaveral locks showed that over the study period, the river on an average annual basis was at or above 1.0 foot NGVD seventy-three days a year. Finally, Mr. Lofton, using water level data gathered over a period of seventeen months at the culvert and the water level data gathered at the Port Canaveral locks, performed a regression analysis. See Government exhibit 56. This analysis purported to show a strong correlation between the water level at the locks and the culvert.
*370The hydrologist’s conclusions, however, are not viewed by the court as conclusive evidence that water levels at or above .9 feet NGVD at the Port Canaveral locks cause innundation on the defendants’ wetlands. The regression analysis did not take into account independent factors contributing to the water level at the site such as well water flowing from the defendants’ property into the culvert or surface water flowing across the defendants’ property from the north from a disposal site which the evidence indicates exists to the north and west of the defendants' property. Furthermore, water levels at the culvert do not necessarily predict an equal water level on the defendants’ wetlands. Mr. Lofton admitted that a higher water level in the culvert was required to connect the water to the eastern wetland than to the central or western wetland. He testified that it would require a water level between 1.0 foot and 1.1 feet NGVD to connect water in the culvert with the eastern wetland.
As part of his investigation, Dr. Banner studied soil samples from the wetlands areas. He determined that soils in the wetland areas had three to six inches of organic material at the surface. Based on his studies, Dr. Banner concluded that the defendants’ property contained prior to filling operations three “wetland areas” as that term is defined by the Army Corps of Engineers regulation found in 33 C.F.R. 323.2(c) (1980). The factors considered by Dr. Banner were summarized for the court’s benefit on plaintiff’s exhibit 10. The wetlands found by Dr. Banner were delineated on a survey map in evidence as Government exhibit 20. Although Dr. Banner’s wetland determination is contradicted in some aspects by respectable expert witnesses presented by the defendants, it is also corroborated in some respects by those and by other witnesses, particularly Conrad White, Irene Sadowski, and William Kruczynski. On the basis of Dr. Banner’s testimony, the court finds that the Government has proved by a preponderance of the evidence that the defendants’ property before the filling operations of Mr. Lambert, contained three wetland fingers and a wetland bridge between the western wetland and the central wetland as depicted on Government exhibit 20.
In reaching this finding the court has concluded that proof by a preponderance of the evidence is the burden of persuasion upon the Government. See United States v. Ward, 448 U.S. 242, 248, 100 S.Ct. 2636, 2641, 65 L.Ed.2d 742 (1980), and United States v. Regan, 232 U.S. 37, 47, 34 S.Ct. 213, 216, 58 L.Ed. 494 (1914). The court, however, has also concluded that in assessing the weight to be given the testimony of the various witnesses, the court must give substantial deference to the well reasoned conclusions of those Government witnesses who are officials charged by law with administering the provisions of the Clean Water Act. See National Wildlife Federation v. Marsh, 721 F.2d 767 (11th Cir.1983), and Avoyelles Sportsmen’s League, Inc. v. Marsh, 715 F.2d 897 (5th Cir.1983).
In Avoyelles, the case was before the district court in a different posture than is the present. In Avoyelles, the trial judge had before him a formal agency determination of wetland boundaries before he was called upon to rule on that same issue. In that circumstance, the trial court was limited in its role to determination of whether or not the agency finding was arbitrary and capricious. In the present casé, no formal agency record was established and presented to this court for review; therefore, this court has assumed a broader latitude in dealing with the position of the Government, although it is still bound as stated above to give significant deference to the testimony of responsible Government officials.
There are other reasons why the court has determined to accept as reliable the conclusion of Dr. Banner as to the extent of wetlands on defendants’ property. The defendants' principal expert witness, Dr. George Cornwell, conceded that the defendants’ property had a central and western wetland and, to that extent, he corroborated the testimony of Dr. Banner. Dr. *371Cornwell’s main objection to the characterization of the eastern finger as wetland was to the use of plants tolerant to life in saturated soil as indicators of wetland. He testified that such vegetation was not “typically adapted” to life in saturated soil. He testified that if a plant lives in wetlands, but also lives in dry areas, it is “a lousy indicator” of wetlands. Regardless of the logic of this conclusion, it has been held that the Corps regulations may encompass plants which in fact are tolerant to life in saturated soil even though they may not live out an entire life cycle in such soil. Avoyelles Sportsmen’s League, Inc. v. Marsh, supra, at 913.
The evidence before the court indicates that there is a close physical proximity between the defendants’ wetlands as defined by Dr. Banner and the Banana River. The parties agree that the Banana River is a navigable water. Additionally, both defense and Government witnesses agree that there is a ground water connection between the Banana River and the defendants’ property. A ground water connection to navigable water causing saturation has been held sufficient to bring property within the Corps’ wetlands definition provided the other elements of the definition are present. See United States v. Tilton, 705 F.2d 429 (11th Cir.1983), note 1 (but for an apparently contrary holding see United States v. Riverside Bayview Homes, Inc. et al., 729 F.2d 391 (6th Cir.1984).
Although the extent of surface water intrusion on defendants’ property from the Banana River is in sharp dispute, it is clear from the evidence that the culvert through Banana River Drive provides a means for such intrusion. It is also indisputable that on some occasions water from the river flows through the culvert and onto defendants’ wetlands. Further evidence of the adjacency of defendants’ wetlands to the Banana River was provided by the testimony of Dr. William Kruczynski, a Ph.D. employed by the Environmental Protection Agency. He characterized the upland areas on defendants’ property and the road lying east of the property as “relic dunes” separating the wetlands from the Banana River. Based on the foregoing evidence, the court finds that the defendants’ wetlands are “adjacent” for purposes of Section 323.2(a)(2), 33 C.F.R. (1980).
There is no substantial dispute that the fill placed in the wetlands was a “discharge of a pollutant” within the purview of 33 U.S.C. § 1311(a) and 1362(12). The defendant argues that his filling operations were permitted under Section 323.4-2 of the Corps regulations as they existed in 1980. 33 C.F.R. § 323.4-2(a)(l) (1980) reads as follows:
(a) Discharges of dredged or fill material into the following waters of the United States are hereby permitted for purposes of Section 404, provided the conditions in paragraph (b) below are met:
(1) Non-tidal rivers, streams and their impoundments including adjacent wetlands that are located above the headwaters;
It is, however, inapplicable. The defendants’ wetlands are adjacent to the Banana River which even if non-tidal has not been shown to have a flow less than five cubic feet per minute. The fill, therefore, was not placed above “headwaters” as defined in 33 C.F.R. § 323.2(i) (1980). Hence the court concludes that the placing of fill in the wetlands as depicted on Government Exhibit 21 by the defendant William Lambert constituted a series of violations of Section 1311(a) of Title 33, United States Code. These violations began in or about 1977 and ended in or about March 1982.
Restoration
Under Section 1319(b) and (d), Title 33, United States Code, the court has the authority to compel restoration and impose a civil penalty as a means of enforcing compliance with Section 1311 of Title 33. A violation of Section 1311, however, does not automatically require restoration. See United States v. Context-Marks Corp., 729 F.2d 1294 (11th Cir.1984). In considering the restoration issue, the three wetlands fingers should be considered separately.
*372The eastern wetland was shown.by the evidence to have been a transitional area, that is, an area transitioning to upland status. Mr. White characterized the eastern wetland as a “stressed” wetland producing a number of plants not indicative of a wetland. William Kruczynski, a Government rebuttal witness, testified that the construction of State Road 528 (shown on Government exhibit 112) has contributed to the drying out process on the eastern wetland. One of the defendants’ experts, Mr. J.B. Butler, testified that the natural watershed to the defendants’ property was severely truncated by the construction of State Road 528 (also referred to as Bennett Causeway). Mr. Butler testified that seventy-five to eighty percent of the eastern wetland was over one foot NGVD before the defendants’ filling commenced. It was not until after the present litigation commenced that any Government agent classified the north half of the eastern wetland as a wetland area within the Corps jurisdiction. The foregoing evidence and the lack of a significant and clearly demonstrated surface water connection between the eastern wetland and the Banana River lead the court to conclude that restoration of the eastern wetland would be an unduly burdensome and inequitable application of injunctive power.
The Government has urged (he court to order restoration of the, eastern wetland to prevent Mr. Lambert from profiting through the fill operation in violation of the Clean Water Act. The Government’s factual basis for this argument is the testimony of Philip Pickens, a real estate appraiser. Mr. Pickens testified that Mr. Lambert had considerably enhanced the value of his property by the filling. There are two reasons why the Government’s position should be rejected. First, Mr. Pickens’ testimony dealing with enhancement is subject to question on a competency basis. Mr. Pickens, although a qualified appraiser, made certain assumptions which render the probative value of his opinion questionable. He testified that nine acres on Mr. Lambert’s property were filled and assumed that the fill would permit commercial use of the land. There was, however, no factual predicate for this assumption. Furthermore, he admitted that he did not go on the property for purposes of a site examination and did not know if the filled areas had a sufficient configuration to permit use for anything. Given these limitations on Mr. Pickens’ testimony, the court cannot determine the extent to which, if at all, the fill enhanced the value of the Lambert site. Additionally, the injunctive power possessed by the court under the Clean Water Act is not designed to penalize a landowner for improving his property. It is designed to permit the court to restore favorable environmental factors which have been disrupted by unlawful filling activities. This court concludes that the evidence is not sufficiently clear that the filling of the eastern wetland so significantly disrupted the navigable body of water to which it is adjacent that an order of restoration would be consistent with the burden which would thereby be imposed on the landowner.
The court has also not overlooked the testimony of Dr. William Kruczynski to the effect that he saw highly organic water coming from the fill on the eastern wetland at a location on the southeast corner of the defendants’ property and moving from there into the roadside ditch which leads to the culvert and ultimately to the Banana River. This testimony, however, is not sufficiently strong to require the court to order restoration of the entire eastern wetland. The probative value of this testimony was significantly undermined on cross-examination which elicited an admission from the witness that in previous deposition testimony he placed the site of the leachate on the east side of the defendants’ property near a gate which is north of the defendants’ southeast corner and not proximately related to the ditch. The testimony of Mr. Delbert Hicks, an expert in aquatic biology with the EPA, also suggests that un-ionized ammonia may be forming in the fill area and could move thence to the Banana River via the culvert. However, the amount of such ammonia was not quanti*373fied and a conclusion on the present record that it presents a threat to the ecology of the Banana River is too speculative to justify removal of the fill from the eastern wetland.
With regard to the central wetland, the evidence shows that the defendant by excavation created a six acre lake. The excavation of the lake was not a violation of the Clean Water Act as was earlier noted by this court’s opinion in this case in connection with the denial of the preliminary injunction. In connection with the excavation of the lake, fill pads were placed in the central wetland in violation of the Act. These fill pads, however, have now been removed. The only other violation of the Act which occurred in connection with the excavation of the lake was a de minimis addition of fill to the areas adjacent to the excavation site by drippings from the drag-line as the excavated materials were extracted and placed in trucks. The court, therefore, concludes that there is no justifiable basis under the law for requiring the landowner to fill the excavation which has been referred to in the testimony variously as “Lake Lambert” or “the pit”, depending on one’s point of view.
Mr. Delbert Hicks testified that ammonia from the pit “might” move out of the pit and into the roadside ditch and thence, via the culvert, to the Banana River. The ammonia is forming by a combination of water in the pit with organic nitrogen leaching into the pit from the filled areas.
Mr. Hicks could not say how long the leachate will continue to produce ammonia in the pit. He testified that only ammonia in un-ionized form is harmful to marine life. At concentrations of .34 to 3.2 milligrams per liter, un-ionized ammonia is lethal and at .03 to .12 milligrams per liter, un-ionized ammonia is debilitating to certain marine species. His testimony with regard to the ammonia was based in part upon a measurement of ammonia level at the culvert in September of 1983. The measurement at that time produced a finding of .4 milligrams per liter of un-ionized ammonia. Mr. Hicks, however, was totally unaware when giving this testimony of any source producing ammonia at the culvert other than the Lambert pit. Specifically he was unaware of a source of ammonia northwest of the Lambert site. Later testimony by Mr. Lambert indicated that a fill area exists northwest of his property where scallop shells are being used for fill. Mr. Hicks admitted on rebuttal that water coming from that fill area northwest of the Lambert property is a source of ammonia to the culvert. Furthermore, Mr. Hicks admitted that the ammonia forming in Lake Lambert is forming only at depths where there is no oxygen. In the area of the lake nearest to the mechanical connection provided by two PVC pipes running from the south end of the lake, no anoxic condition was found. In other areas of the lake to the north of its southern' extreme, anoxic conditions were found at depths between 8 and 11 feet.
Dr. Joseph Delfino, a water quality expert for the defendants, testified that it was not possible that the site could be affecting the ammonia and nitrogen content of the Banana River. He testified that the source of ammonia northwest of the Lamberts’ property could account for some of the un-ionized ammonia found by the EPA at the culvert. He further testified that the Government data on the quantity of hydrogen sulfide and ammonia was insufficient to quantify either the un-ionized hydrogen sulfide or ammonia in the lake. He also testified that the water quality in Lake Lambert is typical of that in the surrounding areas.
Because the excavation of the lake with minor exceptions did not involve an illegal fill and the water of the lake has not been proved by a preponderance of the evidence to be itself a pollutant of the Banana River, the court concludes there is no justifiable basis for ordering a fill of the lake. As an incidental matter, it should be noted the pit on the Lambert property was not alleged in the complaint to be a point source of pollution and its status as such is actually beyond the issues framed by the complaint.
*374With regard to the western wetland which all parties conclude is the wettest of the areas on the defendants’ property, the court finds that removal of the fill, including the fill placed in the bridge area, would not be unduly burdensome to the defendants and would significantly protect and preserve the value of the western wetland as a contributing factor in the environment surrounding the Banana River. The parties are, therefore, directed to submit to the court within thirty days from the filing of this opinion a proposed final judgment and a stipulated plan for the restoration of the western wetland. If the parties are not able to stipulate on a restoration plan, they shall within the thirty day period submit separate plans for restoration. Any such plan shall specify a time for the commencement of the restoration work and also a maximum time within which the work is to be concluded. The plan shall also provide for periodic and final inspection of the restoration by a representative of the Environmental Protection Agency.
The court concludes that a civil penalty should be imposed on the defendant William Lambert for the filling of the three wetland fingers in violation of the Clean Water Act. The court will, however, reserve ruling on the amount of the penalty until the court has examined and approved a proposed restoration plan. Additionally the final judgment shall permanently enjoin the defendants from further filling or otherwise discharging a pollutant into the remaining and restored wetlands.
The evidence will not support a finding that Mrs. Lucille H. Lambert actively directed or caused the illegal fill. The court, therefore, concludes that it would be inappropriate to impose a civil penalty on Mrs. Lambert. The evidence does indicate, however, that she was aware that the fill activities were progressing, was aware that they were progressing in areas claimed by the United States Army Corps of Engineers to be wetlands without a permit, and was aware that her husband was responsible for the activities on property jointly owned by her with her husband.
The court, therefore, concludes that the fine which will hereafter be imposed shall constitute a lien on the joint interest of Mr. and Mrs. Lambert in the property and as a joint owner of the property, Mrs. Lambert shall be equally subject to the court’s permanent injunction.