Opinion by Judge LAY; Dissent by Judge GRABER.
LAY, Circuit Judge:Frank and Mary McHugh filed a complaint for a declaratory judgment against the United Service Automobile Association insurance company (“USAA”) and the Director of the Federal Emergency Management Agency (“FEMA”) alleging that (1) their beach house which was insured by USAA under a Standard Flood Insurance Policy (“SFIP”) was damaged by a mudslide, (2) they made a claim to USAA under the SFIP, and (3) USAA improperly denied their claim. The district court granted summary judgment for USAA, holding that the damage to the McHughs’ home was caused by a landslide which is not covered by the policy. We reverse the decision of the district court.
BACKGROUND
Under the National Flood Insurance Act of 1968 (“Act”), codified at 42 U.S.C. §§ 4001 et seq. (1994), the FEMA is authorized to provide federally subsidized flood insurance to individual homeowners. The language of the SFIP is prescribed by the Act and FEMA regulations, although the policy itself is issued through a private insurer. Frank and Mary McHugh purchased a SFIP from USAA to provide coverage for their beach house located on the Hood Canal in Seabeck, Washington.
The McHughs’ Flood Policy defines “Direct Physical Loss By or From Flood” as “any loss in the nature of actual loss of or physical damage, evidenced by physical changes, to the insured property ... which is directly and proximately caused by a ‘flood’ (as defined in this policy).” ER at 31. The policy also defines “Flood” as:
A. A general and temporary condition of partial or complete inundation of normally dry land area from:
1. The overflow of inland or tidal waters.
2. The unusual and rapid accumulation or runoff of surface waters from any source.
3. Mudslides (i.e., mudflows) which are proximately caused by flooding as defined in subparagraph A-2 above and are akin to a river of liquid and flowing mud on the surfaces of normally dry land areas, including your premises, as when earth is carried by a current of water and deposited along the path of the current.
B. The collapse or subsidence of land along the shore of a lake or other body of water as a result of erosion or undermining caused by waves or currents of water exceeding the cyclical levels which result in flooding as defined in subparagraph A-l above.
ER at 31-32 (emphasis added).
The Flood Policy also contains the following exclusion:
We only provide coverage for direct physical loss by or from flood which means we do not cover:
B. Losses from other casualties, including loss caused by:
1. Theft, fire, windstorm, wind, explosion, earthquake, land sinkage, landslide, destabilization or movement of *453land resulting from the accumulation of water in subsurface land areas, gradual erosion, or any other earth movement except such mudslides (i.e., mudflows) or erosion as is covered under the peril of .flood.
ER at 32-33 (emphasis added).
The federal courts have stated that flood insurance polices normally are subject to limitations on coverage imposed by applicable federal regulations. See Carneiro Da Cunha v. Standard Fire Ins. Co./Aetna Flood Ins. Program, 129 F.3d 581, 585 (11th Cir.1997) (stating that “[tjhere is no dispute that the policies at issue are subject to limitations on coverage imposed by federal statute and regulations”); Criger v. Becton, 902 F.2d 1348, 1351 (8th Cir.1990) (stating that FEMA’s interpretation of its own regulations regarding coverage is entitled to great deference). However, this makes little difference in the instant case because the definitions of the terms “flood” and “flooding” in the policy are the same in all material respects as the definitions found in the regulations. The regulations defining mudslide state:
Mudslide (ie., mudflow) describes a condition where there is a river, flow or inundation of liquid mud down a hillside usually as a result of a dual condition of loss of brush cover, and the subsequent accumulation of water on the ground preceded by a period of unusually heavy or sustained rain. A mudslide (i.e., mudflow) may occur as a distinct phenomenon while a landslide is in progress, and will be recognized as such by the Administrator only if the mudflow, and not the landslide, is the proximate cause of damage that occurs.
44 C.F.R. § 59.1 (1997).
On December 28, 1994, the McHughs reported to USAA that their beach house had been damaged by a flood-related mudslide that occurred after heavy rains and an overflow of a drainage ditch situated at the top of the slope on which the house was located. The house, located at the base of a steep sloping hill, was knocked off its foundation and partially destroyed after being hit by a saturated mixture of soil, gravel, vegetation, and rock. After receiving the McHughs’ claim under the SFIP, USAA hired an independent adjustor, who took several photographs,1 and then hired Martin Page from the geo-technical engineering firm of Shannon & Wilson to determine the exact cause of the damage to the McHughs’ house. In a written report, Page concluded that the slide was caused by a combination of factors including saturation of the surface soils by heavy rainfall and a build-up of groundwater in the underlying sand and gravel. Page’s report concludes, however, that “it is our opinion that the soil instability that occurred at the subject site is classified as a landslide, not a mudslide.” ER at 16.
The McHughs hired their own geo-techni-eal engineer, Robert Pride, to assess the cause of the damage. Contrary to Page’s report, Pride found that the damage to the beach house was caused by a mudslide, not a landslide, precipitated by soil saturation and surface-water runoff from a natural drainage channel above the slide area.
In December 1995, the McHughs filed a complaint against USAA and the director of FEMA seeking coverage for damage to the beach house.2 USAA moved for summary judgment in November 1996 to dismiss all claims with prejudice. The McHughs filed a cross-motion for summary judgment a month later asking the district court to hold that their claim was covered by the policy as a matter of law.
On December 6, 1996, the district court3 granted USAA’s motion for summary judgment, holding the damage to the McHughs’ home was caused by a landslide, not a mudslide, and therefore was excluded from coverage under the SFIP. See McHugh v. United Service Automobile Ass’n, No. C95-5702JKA *454(W.D.Wash. Dec. 6, 1996) (Order Granting Defendant’s Motion for Summary Judgment). This appeal followed.
DISCUSSION
The law is clear that, as contracts, SFIPs issued under the National Flood Insurance Program (“NFIP”) are governed by federal law applying standard insurance law principles. See, e.g., Brazil v. Giuffrida, 763 F.2d 1072, 1075 (9th Cir.1985); Carneiro Da Cunha, 129 F.3d at 584. Federal common law therefore controls the interpretation of these insurance policies. Sodowski v. Nat’l Flood Ins. Program of Federal Emergency Management Agency, 834 F.2d 653, 655 (7th Cir.1987). Furthermore, the interpretation of the insurance policy is a question of law for the court and is reviewed de novo. See Stanford Ranch, Inc. v. Maryland Cas. Co., 89 F.3d 618, 624 (9th Cir.1996) (stating that “[t]he interpretation of an insurance policy is a question of law”); Sodowski, 834 F.2d at 655 (analyzing the application of the SFIP as a matter of law on de novo review); Brazil 763 F.2d at 1075 (stating that a reviewing court reviews de novo the district court’s interpretation of an insurance policy).
There were two diverging expert testimonies in this case. Although experts may disagree in their conclusions, their testimony cannot be used to provide legal meaning or interpret the policies as written. See Crow Tribe of Indians v. Racicot, 87 F.3d 1039, 1045 (9th Cir.1996) (stating that expert testimony is not proper for issues of law because the role of experts is to interpret and analyze factual evidence and not to testify about the law); Maffei v. Northern Ins. Co. of New York, 12 F.3d 892, 898-99 (9th Cir.1993) (holding that an insurance expert’s declaration that sulphur dioxide cloud constituted a “hostile fire” as described in insured’s policies was improper expert testimony); Aguilar v. Int’l Longshoremen’s Union Local No. 10, 966 F.2d 443, 447 (9th Cir.1992) (stating that matters of law are “inappropriate subjects for expert testimony”). Therefore, we view the experts’ testimony in this case as only relevant for the historical facts that they observed and not for their legal conclusions as to what conditions were covered or excluded under the terms of the policy.
The historical facts of this case are not in dispute. Robert Pride, the McHughs’ expert, characterized the event as a “saturated soil mass that flowed down towards the [McHughs’] house.” ER at 138. He stated that this signified to him the fluid movement of saturated soils, whatever their makeup. ER at 125-26. The insurance company’s expert did not differ from this observation. In fact, Martin Page, the defendant’s geo-technical engineer, also characterized the event as “a saturated mass of soil and trees” that slid down the slope. ER at 16-17. Page stated in his report:
The slide debris generally consists of gravel and sand mixed with surface vegetation, several large tree stumps, and previously cut tree tops. In our opinion, the presence of cut trees and decayed tree stumps on the surface of the slope may also have contributed to the instability.... The soils that have accumulated against the side of the beach house appear to have slid down the slope as a saturated mass of soil and trees.
ER at 16-17 (emphasis added).4
Pride also observed that the “slide was not a ‘landslide’ ” because the “slide did not involve a deep-seated earth movement or other geo-technical conditions associated with landslides.” ER at 116.5 Furthermore, in his *455report, Pride stated “[l]andslides are typically in excess of five to fifteen feet deep, usually involve large quantities of earth or rock, and may move downslope with much of the slide mass relatively intact.” ER at 22.
It is a fundamental principle that unless the contract terms are specifically different than the common usage of the terms, that the common usage of the terms will be adopted. For example, under the SFIP, a “landslide” as opposed to a mudflow is excluded from coverage. Yet the term landslide is not otherwise defined in the policy. Webster’s dictionary defines a landslide as “the rapid downward movement under the influence of gravity of a mass of rock, earth, or artificial fill on a slope.” Webster’s Third New International Dictionary 1269 (1981).
However, the fundamental question involved here is not so much the interpretation of the exclusionary clause but whether a mudflow occurred and was the proximate cause of the damage to the house. The definition of “mudflow” as set forth in the policy and regulation is reinforced by the common usage of that term. According to Webster’s Ninth New Collegiate Dictionary 778 (1984), “mudflow” is defined as “a moving mass of soil made fluid by rain or melting snow.” We find nothing in the policy itself or in the regulations which defines “mudflow” any differently than the common usage of the term. Under the common usage, the “saturated soil mass” that destroyed the McHughs’ house was a mudslide. •
The dissent faults our interpretation of the record on the basis that there is no evidence offered by the McHughs to show that there was a “liquid river of mud.”6 However, we think that common sense and common meaning must prevail. Under the terms of the policy, “liquid and flowing mud” surely means nothing more than a saturated soil mass moving by liquidity down a slope. The torrential rain and the overflow of the natural drainage channel formed a liquidity which flowed down the slope and brought with it a saturated soil mass composed of soil, sand, gravel, and underbrush. As a result, the McHughs’ house was hit by this saturated mixture of soil, gravel, vegetation and rock, and was knocked off its foundation. The policy itself describes a mudflow “as when earth is carried by a current of water and deposited along the path of the current.” This statement is no different than the definition of the mudflow quoted from Webster’s that it is “a moving mass of soil made fluid by rain or melting snow.”
The district court found that “soil saturation was the predominant cause” of the event which occurred. In denying coverage, however, the district court relied upon Wagner v. Director, Federal Emergency Management Agency, 847 F.2d 515, 521 (9th Cir.1988). The Wagner case is clearly distinguishable from the facts adduced in this ease. First and foremost, as the Wagner court points out, in that case the plaintiffs admitted that floodwaters did not damage their property directly and all of their losses were caused by the shifting of the saturated earth beneath their home. Unlike the facts in Wagner, the McHughs’ home was hit by the liquid flow of soil mass which in turn knocked the home off its foundation. There. is no evidence that the saturation of the land underneath the McHughs’ home gave way or *456was caused to move by reason of that saturation. The pictures submitted to the court by exhibit clearly demonstrate to the contrary. Neither expert testified that saturation of the land underneath the house caused the movement of the house.
In the Wagner case, the court stated that “the' courts have all but universally held, federal flood insurance policies do not cover losses stemming from water-caused earth movements.” Wagner, 847 F.2d at 522. It is clear from the facts involved and the cases cited to support this statement, that the court did not mean to exclude water-caused mudflows. In other words, Wagner’s statement was not intended to change the policy terms which define mudflow “as when earth is carried by a current of water____” ER at 32.
The Wagner court finds support for its statement in the circuit court opinions of Sodowski v. National Flood Ins. Program of Federal Emergency Management Agency, 834 F.2d 653, 657-59 (7th Cir.1987),7 and West v. Harris, 573 F.2d 873 (5th Cir.1978). However, the facts in Sodowski and West demonstrate that these cases are completely distinguishable from the present case. Neither Sodowski or West involved a soil mass flowing down a hill caused by heavy rains and overflow of a drainage ditch. As the court pointed out in Sodowski:
In West, the court was faced with two similar factual situations, both involving houses built on concrete slabs. There, as in this case, flooding occurred in the area surrounding the two dwellings, and in one instance water also inundated the house. The soil under the insureds’ structures became saturated with water and eventually the soil settled causing structural damage to the dwelling. The court stated that “[r]egardless of whether this settlement had been in process over a long period of time or whether it occurred immediately after the flood and draining of the canals, it was still the result of earth movement.” West v. Harris, 573 F.2d at 877. The court, giving full effect to the earth movement exclusion and disallowing the plaintiffs’ claims, held that “[t]he policy does not cover loss caused by earth movement in the form of soil settlement.” Id. [footnote omitted].
Sodowski, 834 F.2d at 656-57.
The court in Wagner also compared the case to Atlas Pallet, Inc. v. Gallagher, 725 F.2d 131, 137 (1st Cir.1984), which held that a SFIP does not cover damage caused by a flood induced mill dam collapse. A reading of the Gallagher case, however, clearly shows that the issue in that case was whether the mill dam was a structure covered by the policy. There is no question of similar issues of coverage in the present case.
Under the reading of the policy urged by USSA, a policyholder whose home was inundated with flood waters mixed with fine, granular soil would receive coverage, while a homeowner unfortunate enough to be located below a slope with sand and gravel where flood waters are mixed with fine, granular soil would be denied coverage. While FEMA should be free to draw a mudslide/Iandslide distinction when allocating its flood coverage, as a matter of fairness such a distinction should not lead to such a seemingly arbitrary result.
The Act mandates that coverage under the term “flood” includes “inundation from mudslides which are proximately caused by accumulations' of water on or under the ground. ...” 42 U.S.C. § 4121(b) (1994) (emphasis added). See also 42 U.S.C. § 4001(f) (1994) (providing “It is therefore the further purpose of this chapter to make available ... protection against damage and loss resulting from mudslides that are caused by accumulations of water on or under the ground.”) (emphasis added). Even the district court’s conclusion that soil saturation was the predominant cause of the damage at the McHughs’ property fits within the partial definition of a mudslide proximately caused by accumulations of water on or under the ground.
*457CONCLUSION
As we earlier observed, whether coverage exists under an insurance policy is a question of- law. The historical facts in the present case are not materially disputed by the experts’ reports and depositions. What is disputed is their legal conclusions as to whether the facts fall within the terms of “mudflow” or the exclusion of a “landslide” within the policy. This is a question of law for the court to be decided from the definitions within the policy as reinforced by the common usage of the terms. The following facts are undisputed:
(1) At the time of the damage to McHughs’ home, there was a heavy rainfall and overflow of waters from a drainage ditch located above the house and slope.
(2) The waters came down the slope, inundated the slope, and carried a mass of saturated soil, gravel, vegetation and rock.
(3) The saturated soil mass flowed down the slope, hit the McHughs’ home, and knocked it off its foundation. (See Appendix A).
(4) The damage to the foundation did not occur because of settlement of saturated soil underneath the house. Cf. Wagner and Sodowski, supra.
We conclude that the policy is not ambiguous in its terms. We also conclude that the occurrence was “akin to a river of liquid, flowing mud” and constituted “earth carried by a current of water and deposited along the path of the current.” These definitions simply mean that coverage will be extended to the common usage of the term “mudflow” as when damage is proximately caused “by a moving mass of soil made fluid by rain.”
We, therefore, conclude that under the undisputed historical facts the coverage of the SFIP policy clearly extends to the flooding (i.e., mudflow) that proximately caused the damage to the McHughs’ home. We vacate the judgment of the district court and remand the ease for entry of judgment in favor of the plaintiffs.
. We have attached as an Appendix to the opinion four of those photographs which visibly demonstrate part of the debris and saturated soil mass which pushed the home off its foundation.
. The director of FEMA was dismissed as a party on March 11, 1996.
.The parties consented for U.S. Magistrate Judge Kelley Arnold for the Western District of Washington, to hear and decide the cross-motions for summary judgment.
. The experts, however, disagreed about whether mud flowed around the house. During Pride's deposition, he was asked whether the debris flowed around the sides of the house. In response, Mr. Pride stated that "at the time that I was there, no, the material had been removed away from the north side. Frank McHugh had indicated there was some materials that had been taken away to gain access, but how much of it, I didn’t see.” ER at 124.
This contradicted Page’s statement, "There was no evidence of soils having flowed around the sides of or into the house, as would have occurred if there had been significant flowing of wet or saturated soils.” ER at 17.
. In his initial report, Pride, stated that the saturated soil mass was a "sand-gravel flow,” not a landslide:
Once saturated, the outer weathered zone gained weight and lost strength, resulting in the soil, rock and debris flow. If the saturated soils on the slope consisted of silts and clays, *455or even sands and silts, the failure most certainly would have been labeled a "mudflow.” Saturated fine-grained soils would behave more like a slurry because of its particle size, strength and viscosity characteristics. In contrast, these soils on your slope are basically coarse-grained granular materials with higher strength and greater resistance to failure-thereby resulting in steeper natural slopes. The causes of failure are the same (i.e.: rainfall, surface runoff and groundwater seepage), but the appearance is different. Saturated sands and gravels are not carried as far by water flow, nor do they create mud or slurry-like consistencies. Although excess surface water runoff is more readily absorbed by the native granular soils, the net effect of saturated soils on your steep slope is a "sand-gravel flow” instead of a "mudflow.”
ER at 21.
In his deposition, Pride explained that a "sand-gravel flow” would be included within "the generic sense of mudflow." ER at 32. As we discuss, however, neither expert was qualified to give an opinion as to whether the occurrence was covered or excluded under the terms of the policy.
. The regulation reads "a river, flow or inundation of liquid mud ...” 44 C.F.R. § 59.1 (1997).
. The SFIP policy issued to the plaintiff in these cases was worded differently than the McHughs' current policy.